CCIE routing and switching certification guide [4th ed] 9781587059803, 1587059800

Table of contents :
Cover......Page 1
About the Authors......Page 5
Acknowledgments......Page 6
Contents......Page 10
Icons Used in This Book......Page 30
Command Syntax Conventions......Page 31
Foreword......Page 32
Introduction......Page 33
“Do I Know This Already?” Quiz......Page 48
Ethernet Layer 1: Wiring, Speed, and Duplex......Page 52
Ethernet Layer 2: Framing and Addressing......Page 58
Switching and Bridging Logic......Page 63
SPAN and RSPAN......Page 65
Foundation Summary......Page 71
Further Reading......Page 74
“Do I Know This Already?” Quiz......Page 76
Virtual LANs......Page 80
VLAN Trunking Protocol......Page 87
VLAN Trunking: ISL and 802.1Q......Page 93
Configuring PPPoE......Page 101
Foundation Summary......Page 104
Memory Builders......Page 105
Further Reading......Page 106
“Do I Know This Already?” Quiz......Page 108
802.1d Spanning Tree Protocol......Page 112
Optimizing Spanning Tree......Page 124
Protecting STP......Page 133
Troubleshooting Complex Layer 2 Issues......Page 136
Foundation Summary......Page 146
Further Reading......Page 148
“Do I Know This Already?” Quiz......Page 150
IP Addressing and Subnetting......Page 153
CIDR, Private Addresses, and NAT......Page 170
Foundation Summary......Page 180
Fill in Key Tables from Memory......Page 183
Further Reading......Page 184
“Do I Know This Already?” Quiz......Page 186
ARP, Proxy ARP, Reverse ARP, BOOTP, and DHCP......Page 191
HSRP, VRRP, and GLBP......Page 195
Network Time Protocol......Page 199
SNMP......Page 200
Syslog......Page 204
Web Cache Communication Protocol......Page 205
Foundation Summary......Page 219
Further Reading......Page 224
“Do I Know This Already?” Quiz......Page 226
IP Forwarding......Page 231
Multilayer Switching......Page 240
Policy Routing......Page 246
Optimized Edge Routing and Performance Routing......Page 251
Foundation Summary......Page 258
Further Reading......Page 260
“Do I Know This Already?” Quiz......Page 262
EIGRP Basics and Steady-State Operation......Page 266
EIGRP Convergence......Page 273
EIGRP Configuration......Page 279
Foundation Summary......Page 289
Definitions......Page 291
Further Reading......Page 292
“Do I Know This Already?” Quiz......Page 294
OSPF Database Exchange......Page 299
OSPF Design and LSAs......Page 314
OSPF Configuration......Page 333
Foundation Summary......Page 347
Memory Builders......Page 351
Further Reading......Page 352
“Do I Know This Already?” Quiz......Page 354
Route Maps, Prefix Lists, and Administrative Distance......Page 359
Route Redistribution......Page 366
Route Summarization......Page 384
Default Routes......Page 387
Troubleshooting Complex Layer 3 Issues......Page 394
Foundation Summary......Page 406
Further Reading......Page 408
“Do I Know This Already?” Quiz......Page 410
Foundation Topics......Page 415
Building BGP Neighbor Relationships......Page 416
Building the BGP Table......Page 425
Building the IP Routing Table......Page 447
Foundation Summary......Page 465
Definitions......Page 469
Further Reading......Page 470
“Do I Know This Already?” Quiz......Page 472
Route Filtering and Route Summarization......Page 478
BGP Path Attributes and the BGP Decision Process......Page 497
Configuring BGP Policies......Page 503
BGP Communities......Page 523
Foundation Summary......Page 531
Further Reading......Page 535
“Do I Know This Already?” Quiz......Page 538
Fields That Can Be Marked for QoS Purposes......Page 542
Cisco Modular QoS CLI......Page 548
Classification and Marking Tools......Page 553
AutoQoS......Page 564
Foundation Summary......Page 569
Definitions......Page 571
Further Reading......Page 572
“Do I Know This Already?” Quiz......Page 574
Cisco Router Queuing Concepts......Page 578
Queuing Tools: CBWFQ and LLQ......Page 580
Weighted Random Early Detection......Page 591
Modified Deficit Round-Robin......Page 595
LAN Switch Congestion Management and Avoidance......Page 597
Resource Reservation Protocol (RSVP)......Page 604
Further Reading......Page 610
“Do I Know This Already?” Quiz......Page 612
Traffic-Shaping Concepts......Page 617
Generic Traffic Shaping......Page 621
Class-Based Shaping......Page 623
Frame Relay Traffic Shaping......Page 629
Policing Concepts and Configuration......Page 635
QoS Troubleshooting and Commands......Page 646
Foundation Summary......Page 651
Definitions......Page 653
Further Reading......Page 654
“Do I Know This Already?” Quiz......Page 656
Point-to-Point Protocol......Page 659
Frame Relay Concepts......Page 667
Frame Relay Configuration......Page 673
Foundation Summary......Page 683
Definitions......Page 686
“Do I Know This Already?” Quiz......Page 688
Why Do You Need Multicasting?......Page 691
Multicast IP Addresses......Page 697
Managing Distribution of Multicast Traffic with IGMP......Page 702
LAN Multicast Optimizations......Page 717
Memory Builders......Page 731
References in This Chapter......Page 732
“Do I Know This Already?” Quiz......Page 734
Multicast Routing Basics......Page 738
Dense-Mode Routing Protocols......Page 745
Sparse-Mode Routing Protocols......Page 762
Foundation Summary......Page 791
Fill In Key Tables from Memory......Page 795
Further Reading......Page 796
“Do I Know This Already?” Quiz......Page 798
Router and Switch Device Security......Page 802
Layer 2 Security......Page 811
Layer 3 Security......Page 828
Foundation Summary......Page 856
Memory Builders......Page 859
Further Reading......Page 860
“Do I Know This Already?” Quiz......Page 862
MPLS Unicast IP Forwarding......Page 866
MPLS VPNs......Page 884
Other MPLS Applications......Page 915
VRF Lite......Page 917
Further Reading......Page 922
“Do I Know This Already?” Quiz......Page 924
Foundation Topics......Page 928
IPv6 Addressing and Address Types......Page 929
Basic IPv6 Functionality Protocols......Page 939
Access Lists......Page 948
IPv6 Static Routes......Page 949
IPv6 Unicast Routing Protocols......Page 951
OSPFv3......Page 952
EIGRP for IPv6......Page 963
Route Redistribution and Filtering......Page 972
Quality of Service......Page 976
Tunneling Techniques......Page 978
IPv6 Multicast......Page 985
Foundation Summary......Page 989
Definitions......Page 991
Further Reading......Page 992
Appendix A: Answers to the “Do I Know This Already?” Quizzes......Page 994
Appendix B: Decimal to Binary Conversion Table......Page 1024
Appendix C: CCIE Exam Updates......Page 1028
A......Page 1031
B......Page 1033
C......Page 1034
D......Page 1039
E......Page 1041
F......Page 1042
H......Page 1043
I......Page 1044
L......Page 1046
M......Page 1047
N......Page 1049
O......Page 1050
P......Page 1051
Q......Page 1053
R......Page 1054
S......Page 1056
T......Page 1059
V......Page 1060
Z......Page 1061
Appendix D: IP Addressing Practice......Page 1069
Appendix E: RIP Version 2......Page 1145
Appendix F: IGMP......Page 1167
Appendix G: Key Tables for CCIE Study......Page 1171
Appendix H: Solutions for Key Tables for CCIE Study......Page 1229
Glossary......Page 1285
A......Page 1286
B......Page 1290
C......Page 1292
D......Page 1297
E......Page 1301
F......Page 1304
G......Page 1306
H......Page 1308
I......Page 1309
L......Page 1312
M......Page 1316
N......Page 1320
O......Page 1323
P......Page 1325
Q......Page 1329
R......Page 1330
S......Page 1335
T......Page 1341
V......Page 1344
W......Page 1346
Z......Page 1347

Citation preview

CCIE Routing and Switching Certification Guide Fourth Edition Wendell Odom, CCIE No. 1624 Rus Healy, CCIE No. 15025 Denise Donohue, CCIE No. 9566

Cisco Press 800 East 96th Street Indianapolis, IN 46240 USA

ii

CCIE Routing and Switching Certification Guide, Fourth Edition Wendell Odom, CCIE No. 1624 Rus Healy, CCIE No. 15025 Denise Donohue, CCIE No. 9566 Copyright © 2010 Pearson Education, Inc. Published by: Cisco Press 800 East 96th Street Indianapolis, IN 46240 USA All rights reserved. No part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or by any information storage and retrieval system, without written permission from the publisher, except for the inclusion of brief quotations in a review. Printed in the United States of America First Printing November 2009 Library of Congress Cataloging-in-Publication Data Odom, Wendell. CCIE routing and switching exam certification guide / Wendell Odom, Rus Healy, Denise Donohue. -- 4th ed. p. cm. Includes index. ISBN-13: 978-1-58705-980-3 (hardcover w/cd) ISBN-10: 1-58705-980-0 (hardcover w/cd) 1. Telecommunications engineers—Certification—Study guides. 2. Routing (Computer network management)—Examinations—Study guides. 3. Telecommunication—Switching systems—Examinations—Study guides. 4. Computer networks—Examinations—Study guides. 5. Internetworking (Telecommunication)—Examinations—Study guides. I. Healy, Rus. II. Donohue, Denise. III. Title. QA76.3.B78475 2010 004.6—dc22 2009041604 ISBN-13: 978-1-58705-980-3 ISBN-10: 1-58705-980-0

Warning and Disclaimer This book is designed to provide information about Cisco CCIE Routing and Switching Written Exam, No. 350-001. Every effort has been made to make this book as complete and as accurate as possible, but no warranty or fitness is implied. The information is provided on an “as is” basis. The authors, Cisco Press, and Cisco Systems, Inc. shall have neither liability nor responsibility to any person or entity with respect to any loss or damages arising from the information contained in this book or from the use of the discs or programs that may accompany it. The opinions expressed in this book belong to the author and are not necessarily those of Cisco Systems, Inc.

Trademark Acknowledgments All terms mentioned in this book that are known to be trademarks or service marks have been appropriately capitalized. Cisco Press or Cisco Systems, Inc., cannot attest to the accuracy of this information. Use of a term in this book should not be regarded as affecting the validity of any trademark or service mark.

iii

Corporate and Government Sales Cisco Press offers excellent discounts on this book when ordered in quantity for bulk purchases or special sales. For more information, please contact: U.S. Corporate and Government Sales 1-800-382-3419 [email protected] For sales outside of the U.S. please contact: International Sales 1-317-581-3793 [email protected]

Feedback Information At Cisco Press, our goal is to create in-depth technical books of the highest quality and value. Each book is crafted with care and precision, undergoing rigorous development that involves the unique expertise of members from the professional technical community. Readers’ feedback is a natural continuation of this process. If you have any comments regarding how we could improve the quality of this book, or otherwise alter it to better suit your needs, you can contact us through email at [email protected]. Please make sure to include the book title and ISBN in your message. We greatly appreciate your assistance. Publisher: Paul Boger Associate Publisher: Dave Dusthimer Cisco Representative: Erik Ullanderson Cisco Press Program Manager: Anand Sundaram Executive Editor: Brett Bartow Managing Editor: Patrick Kanouse Development Editor: Dayna Isley Project Editor: Seth Kerney Copy Editor: Keith Cline Technical Editor(s): Maurilio Gorito, Narbik Kocharians Editorial Assistant: Vanessa Evans Book Designer: Louisa Adair Composition: Mark Shirar Indexer: Tim Wright Proofreader: Apostrophe Editing Services

iv

About the Authors Wendell Odom, CCIE No. 1624, is a 28-year veteran of the networking industry. He currently works as an independent author of Cisco certification resources and occasional instructor of Cisco authorized training for Skyline ATS. He has worked as a network engineer, consultant, systems engineer, instructor, and course developer. He is author of several best-selling Cisco certification titles. He maintains lists of current titles, links to Wendell’s blogs, and other certification resources at www.TheCertZone.com. Rus Healy, CCIE No. 15025, has worked on several Cisco Press projects, including the third edition of this book as a coauthor, and the second edition as a technical reviewer. Rus is chief technology officer of Annese & Associates, Cisco’s Education Partner of the Year for the Northeast US and Canada. Rus serves on the Board of Directors of Habitat for Humanity of New York State and Habitat for Humanity of Ontario County, NY. Denise Donohue, CCIE No. 9566, is senior solutions architect for ePlus Technology, a Cisco Gold partner. She works as a consulting engineer, designing networks for ePlus’s customers. Prior to this role, she was a systems engineer for the data consulting arm of SBC/AT&T. She has co-authored several Cisco Press books in the areas of route/switch and voice. Denise has been a Cisco instructor and course director for Global Knowledge and did network consulting for many years. Her areas of specialization include route/switch, voice, and data center.

About the Technical Reviewers Maurilio Gorito, CCIE No. 3807 (Routing and Switching, WAN Switching, and Security), has more than 20 years of experience in networking, including Cisco networks and IBM/ SNA environments, which includes the planning, designing, implementation, and troubleshooting of large IP networks running RIP, IGRP, EIGRP, BGP, OSPF, QoS, and SNA worldwide, including in Brazil and the United States. Maurilio has worked for Cisco since 2000 with the CCIE Team. As program manager, he is responsible for managing the CCIE Routing and Switching track certification exams, and he has more than seven years of experience proctoring CCIE lab exams. He holds degrees in mathematics and pedagogy. Narbik Kocharians, CCIE No. 12410 (Routing and Switching, Security, SP), is a Triple CCIE with more than 32 years of experience in the IT industry. He has designed, implemented, and supported numerous enterprise networks. Narbik is the president of Micronics Training Inc. (www.Micronicstraining.com), where he teaches CCIE R&S and SP boot camps.

v

Acknowledgments Maurilio and Narbik each did a nice job tech editing the book and finding the technical errors that can creep into a manuscript. On his third time with editing this book, Maurilio did his usual great job with one of the most difficult challenges with this book: help us choose what to cover, and in what depth, and what to not cover. And what a treat to get Narbik, one of the world’s most respected CCIE instructors, to review the book. His comments both on technical accuracy and suggested improvements of how to go about describing the topics were very valuable. Joe Harris (CCIE 6200, R/S, Security, SP) did a great job for us working to update and add to the CD question bank. Joe's expertise and experience has been a tremendous help to improve the questions on the CD. Thanks, Joe! We had the privilege of working with Dayna Isley as development editor this time around. Dayna got the task of juggling a wide variety of details, keeping track of a large number of chapters, some with few changes, some with many small changes, and some with big chunks of new material that needed to fit well with existing material (and with 3 authors to boot). And oh yeah, she had to do the usual development work, too. Amazing job, Dayna! The wonderful (and mostly hidden) production folks—Patrick Kanouse’s group—did their usual great job. When every time you see how they reworded something, or made a figure look better, or catch a problem, it makes me appreciate the kind of team we have at Cisco Press. In particular, thanks to Seth Kerney for managing the production process as Project Editor for the book, and for working through all the competing tasks, large and small changes, and the competing timelines. Many thanks to the entire production team for pulling us through the process and making the book better. From a more strategic perspective, thanks to Brett Bartow, the executive editor for this book. I can remember sitting at a table at the Cisco Networker’s conference back in… 2004 I believe, and talking with Brett about the possibility of rewriting the first edition of this book for what came to be called the second edition. Not only did Brett work hard, and with flexibility, to get me the chance to write this book originally, he has also helped me keep a great group of co-authors engaged with the book to help use keep the book up-to-date on a relatively frequent update cycle. From Wendell Odom: As usual, the timeline for the new edition of this book coincided with a couple of other projects. Yet again, Rus helped beyond compare. Frankly, while I may have written more net pages in this book overall, Rus has become invested in this book, not just in time and effort, but in the amount he cares about this book in the marketplace. Rus’s value to the ongoing shape of this book goes far beyond any particular words or figures printed in the pages.

vi

Denise Donohue joined the team for this fourth edition, making her the fifth co-author to work on various parts of the book. It was great to have a fresh set of eyes looking at the content, and to have an experienced author and respected consultant/instructor work with us was a big help as well. Without Denise, we never would have made the requested due dates—thanks, Denise! Finally, on the personal side, thanks to my wife, Kris, for helping make this work lifestyle happen for me. I truly love to write, and Kris helps make that happen. Thanks, doll! And as always, thanks to my Lord and Savior, Jesus Christ. From Rus Healy: Thanks to Wendell Odom and Denise Donohue for the opportunity to work with them on this book. It’s been a satisfying and enjoyable project. It’s always a pleasure to serve on a great team, and along with the great folks from Cisco Press, this group is one of the best! Finally, I want to thank my wife, Nancy, and our kids, Gwen and Trevor, for putting up with me as I took time away from family life to work on this book. From Denise Donohue: I would like to second all the wonderful things that Wendell said about the Cisco Press staff and our technical reviewers. Authors are but the tip of the iceberg; producing a quality book requires many hands, and we are so very grateful for all the help. A big “thank you” to Wendell for the opportunity to work on this new edition. The subject matter was interesting, and I learned some new things! What more can you ask? He and Rus are so professional in their writing; my future books will be better because of the tips I picked up from them. I promised my Lord and Savior, Jesus Christ, that I’d give him a shout-out in this book. Thanks to Him for all He’s done, including helping me understand how to explain a tough concept or keep motivated to stay inside and write on bright, sunny spring days. Finally, thank you to my husband and children for picking up the slack while I’m writing. Couldn’t have done it without you!

vii

Contents at a Glance Foreword

xxxi

Introduction

xxxii

Part I

LAN Switching

Chapter 1

Ethernet Basics

Chapter 2

Virtual LANs and VLAN Trunking

Chapter 3

Spanning Tree Protocol

Part II

IP

Chapter 4

IP Addressing

Chapter 5

IP Services

Part III

IP Routing

Chapter 6

IP Forwarding (Routing)

Chapter 7

EIGRP

217

Chapter 8

OSPF

249

Chapter 9

IGP Route Redistribution, Route Summarization, Default Routing, and Troubleshooting 309

Chapter 10

Fundamentals of BGP Operations

Chapter 11

BGP Routing Policies

Part IV

QoS

Chapter 12

Classification and Marking

Chapter 13

Congestion Management and Avoidance

Chapter 14

Shaping, Policing, and Link Fragmentation

Part V

Wide-Area Networks

Chapter 15

Wide-Area Networks

3 31

63

105

141

181

365

427

611

493 529 567

viii

Part VI

IP Multicast

Chapter 16

Introduction to IP Multicasting

Chapter 17

IP Multicast Routing

Part VII

Security

Chapter 18

Security

Part VIII

MPLS

Chapter 19

Multiprotocol Label Switching

Part IX

IP Version 6

Chapter 20

P Version 6

Part X

Appendixes

Appendix A

Answers to the “Do I Know This Already?” Quizzes

Appendix B

Decimal to Binary Conversion Table

Appendix C

CCIE Exam Updates

Index

986

643

689

753

817

879

979

983

CD-Only Appendix D

IP Addressing Practice

Appendix E

RIP Version 2

Appendix F

IGMP

Appendix G

Key Tables for CCIE Study

Appendix H

Solutions for Key Tables for CCIE Study

Glossary

949

ix

Contents Foreword

xxxi

Introduction Part I

xxxii

LAN Switching

Chapter 1

Ethernet Basics

3

“Do I Know This Already?” Quiz 3 Foundation Topics 7 Ethernet Layer 1: Wiring, Speed, and Duplex 7 RJ-45 Pinouts and Category 5 Wiring 7 Auto-negotiation, Speed, and Duplex 8 CSMA/CD 9 Collision Domains and Switch Buffering 9 Basic Switch Port Configuration 11 Ethernet Layer 2: Framing and Addressing 13 Types of Ethernet Addresses 15 Ethernet Address Formats 16 Protocol Types and the 802.3 Length Field 17 Switching and Bridging Logic 18 SPAN and RSPAN 20 Core Concepts of SPAN and RSPAN 22 Restrictions and Conditions 22 Basic SPAN Configuration 24 Complex SPAN Configuration 24 RSPAN Configuration 25 Foundation Summary 26 Memory Builders 29 Fill In Key Tables from Memory 29 Definitions 29 Further Reading 29

Chapter 2

Virtual LANs and VLAN Trunking

31

“Do I Know This Already?” Quiz 31 Foundation Topics 35 Virtual LANs 35 VLAN Configuration 35 Using VLAN Database Mode to Create VLANs 36 Using Configuration Mode to Put Interfaces into VLANs Using Configuration Mode to Create VLANs 39 Private VLANs 40 VLAN Trunking Protocol 42 VTP Process and Revision Numbers 43 VTP Configuration 44

38

x

Normal-Range and Extended-Range VLANs Storing VLAN Configuration 47 VLAN Trunking: ISL and 802.1Q 48 ISL and 802.1Q Concepts 48 ISL and 802.1Q Configuration 49 Allowed, Active, and Pruned VLANs 52 Trunk Configuration Compatibility 52 Configuring Trunking on Routers 53 802.1Q-in-Q Tunneling 55 Configuring PPPoE 56 Foundation Summary 59 Memory Builders 60 Fill In Key Tables from Memory 61 Definitions 61 Further Reading 61

Chapter 3

Spanning Tree Protocol

46

63

“Do I Know This Already?” Quiz 63 Foundation Topics 67 802.1d Spanning Tree Protocol 67 Choosing Which Ports Forward: Choosing Root Ports and Designated Ports 67 Electing a Root Switch 67 Determining the Root Port 69 Determining the Designated Port 70 Converging to a New STP Topology 71 Topology Change Notification and Updating the CAM 72 Transitioning from Blocking to Forwarding 73 Per-VLAN Spanning Tree and STP over Trunks 74 STP Configuration and Analysis 76 Optimizing Spanning Tree 79 PortFast, UplinkFast, and BackboneFast 79 PortFast 80 UplinkFast 80 BackboneFast 81 PortFast, UplinkFast, and BackboneFast Configuration 81 PortChannels 82 Load Balancing Across PortChannels 82 PortChannel Discovery and Configuration 83 Rapid Spanning Tree Protocol 84 Rapid Per-VLAN Spanning Tree Plus (RPVST+) 86 Multiple Spanning Trees: IEEE 802.1s 87 Protecting STP 88 Root Guard and BPDU Guard: Protecting Access Ports 89 UDLD and Loop Guard: Protecting Trunks 89

xi

Troubleshooting Complex Layer 2 Issues 91 Layer 2 Troubleshooting Process 91 Layer 2 Protocol Troubleshooting and Commands 92 Troubleshooting Using Basic Interface Statistics 92 Troubleshooting Spanning Tree Protocol 95 Troubleshooting Trunking 95 Troubleshooting VTP 96 Troubleshooting EtherChannels 98 Approaches to Resolving Layer 2 Issues 100 Foundation Summary 101 Memory Builders 103 Fill in Key Tables from Memory 103 Definitions 103 Further Reading 103

Part II IP

Chapter 4

IP Addressing

105

“Do I Know This Already?” Quiz 105 Foundation Topics 108 IP Addressing and Subnetting 108 IP Addressing and Subnetting Review 108 Subnetting a Classful Network Number 109 Comments on Classless Addressing 111 Subnetting Math 111 Dissecting the Component Parts of an IP Address 111 Finding Subnet Numbers and Valid Range of IP Addresses—Binary 112 Decimal Shortcuts to Find the Subnet Number and Valid Range of IP Addresses 113 Determining All Subnets of a Network—Binary 116 Determining All Subnets of a Network—Decimal 118 VLSM Subnet Allocation 119 Route Summarization Concepts 121 Finding Inclusive Summary Routes—Binary 122 Finding Inclusive Summary Routes—Decimal 123 Finding Exclusive Summary Routes—Binary 124 CIDR, Private Addresses, and NAT 125 Classless Interdomain Routing 125 Private Addressing 127 Network Address Translation 127 Static NAT 128 Dynamic NAT Without PAT 130 Overloading NAT with Port Address Translation 131 Dynamic NAT and PAT Configuration 132

xii

Foundation Summary 135 Memory Builders 138 Fill in Key Tables from Memory Definitions 139 Further Reading 139

Chapter 5

IP Services

138

141

“Do I Know This Already?” Quiz 141 Foundation Topics 146 ARP, Proxy ARP, Reverse ARP, BOOTP, and DHCP 146 ARP and Proxy ARP 146 RARP, BOOTP, and DHCP 147 DHCP 148 HSRP, VRRP, and GLBP 150 Network Time Protocol 154 SNMP 155 SNMP Protocol Messages 157 SNMP MIBs 158 SNMP Security 159 Syslog 159 Web Cache Communication Protocol 160 Implementing the Cisco IOS IP Service Level Agreement (IP SLA) Feature Implementing NetFlow 165 Implementing Router IP Traffic Export 166 Implementing Cisco IOS Embedded Event Manager 167 Implementing Remote Monitoring 169 Implementing and Using FTP on a Router 170 Implementing a TFTP Server on a Router 171 Implementing Secure Copy Protocol 171 Implementing HTTP and HTTPS Access 172 Implementing Telnet Access 172 Implementing SSH Access 173 Foundation Summary 174 Memory Builders 179 Fill In Key Tables from Memory 179 Definitions 179 Further Reading 179

Part III IP Routing

Chapter 6

Forwarding (Routing)

181

“Do I Know This Already?” Quiz Foundation Topics 186 IP Forwarding 186

181

163

xiii

Process Switching, Fast Switching, and Cisco Express Forwarding 187 Building Adjacency Information: ARP and Inverse ARP 188 Frame Relay Inverse ARP 189 Static Configuration of Frame Relay Mapping Information 192 Disabling InARP 193 Classless and Classful Routing 194 Multilayer Switching 195 MLS Logic 195 Using Routed Ports and PortChannels with MLS 196 MLS Configuration 197 Policy Routing 201 Optimized Edge Routing and Performance Routing 206 Device Roles in PfR 208 MC High Availability and Failure Considerations 209 PfR Configuration 209 GRE Tunnels 211 Foundation Summary 213 Memory Builders 215 Fill In Key Tables from Memory 215 Definitions 215 Further Reading 215

Chapter 7

EIGRP

217

“Do I Know This Already?” Quiz 217 Foundation Topics 221 EIGRP Basics and Steady-State Operation 221 Hellos, Neighbors, and Adjacencies 221 EIGRP Updates 224 The EIGRP Topology Table 226 EIGRP Convergence 228 Input Events and Local Computation 229 Going Active on a Route 231 Stuck-in-Active 233 Limiting Query Scope 234 EIGRP Configuration 234 EIGRP Configuration Example 234 EIGRP Load Balancing 237 EIGRP Authentication 238 EIGRP Automatic Summarization 239 EIGRP Split Horizon 240 EIGRP Route Filtering 240 EIGRP Offset Lists 242 Clearing the IP Routing Table 243

xiv

Foundation Summary 244 Memory Builders 246 Fill In Key Tables from Memory Definitions 246 Further Reading 247

Chapter 8

OSPF

246

249

“Do I Know This Already?” Quiz 249 Foundation Topics 254 OSPF Database Exchange 254 OSPF Router IDs 254 Becoming Neighbors, Exchanging Databases, and Becoming Adjacent 255 Becoming Neighbors: The Hello Process 257 Flooding LSA Headers to Neighbors 258 Database Descriptor Exchange: Master/Slave Relationship 259 Requesting, Getting, and Acknowledging LSAs 259 Designated Routers on LANs 260 Designated Router Optimization on LANs 260 DR Election on LANs 262 Designated Routers on WANs and OSPF Network Types 263 Caveats Regarding OSPF Network Types over NBMA Networks 264 Example of OSPF Network Types and NBMA 265 SPF Calculation 268 Steady-State Operation 269 OSPF Design and LSAs 269 OSPF Design Terms 270 OSPF Path Selection Process 271 LSA Types and Network Types 271 LSA Types 1 and 2 272 LSA Type 3 and Inter-Area Costs 275 Removing Routes Advertised by Type 3 LSAs 278 LSA Types 4 and 5, and External Route Types 1 and 2 278 OSPF Design in Light of LSA Types 280 Stubby Areas 281 Graceful Restart 284 OSPF Path Choices That Do Not Use Cost 285 Choosing the Best Type of Path 285 Best-Path Side Effects of ABR Loop Prevention 286 OSPF Configuration 288 OSPF Costs and Clearing the OSPF Process 290 Alternatives to the OSPF Network Command 292 OSPF Filtering 293 Filtering Routes Using the distribute-list Command 293

xv

OSPF ABR LSA Type 3 Filtering 295 Filtering Type 3 LSAs with the area range Command Virtual Link Configuration 296 Configuring OSPF Authentication 298 OSPF Stub Router Configuration 301 Foundation Summary 302 Memory Builders 306 Fill In Key Tables from Memory 307 Definitions 307 Further Reading 307

Chapter 9

296

IGP Route Redistribution, Route Summarization, Default Routing, and Troubleshooting 309 “Do I Know This Already?” Quiz 309 Foundation Topics 314 Route Maps, Prefix Lists, and Administrative Distance 314 Configuring Route Maps with the route-map Command 314 Route Map match Commands for Route Redistribution 316 Route Map set Commands for Route Redistribution 317 IP Prefix Lists 318 Administrative Distance 320 Route Redistribution 321 Mechanics of the redistribute Command 321 Redistribution Using Default Settings 322 Setting Metrics, Metric Types, and Tags 325 Redistributing a Subset of Routes Using a Route Map 326 Mutual Redistribution at Multiple Routers 330 Preventing Suboptimal Routes by Setting the Administrative Distance 332 Preventing Suboptimal Routes by Using Route Tags 335 Using Metrics and Metric Types to Influence Redistributed Routes 337 Route Summarization 339 EIGRP Route Summarization 341 OSPF Route Summarization 341 Default Routes 342 Using Static Routes to 0.0.0.0, with redistribute static 344 Using the default-information originate Command 345 Using the ip default-network Command 346 Using Route Summarization to Create Default Routes 347 Troubleshooting Complex Layer 3 Issues 349 Layer 3 Troubleshooting Process 349 Layer 3 Protocol Troubleshooting and Commands 351 IP Routing Processes 352 Approaches to Resolving Layer 3 Issues 359

xvi

Foundation Summary 361 Memory Builders 363 Fill In Key Tables from Memory Definitions 363 Further Reading 363

363

Chapter 10 Fundamentals of BGP Operations

365

“Do I Know This Already?” Quiz 365 Foundation Topics 370 Building BGP Neighbor Relationships 371 Internal BGP Neighbors 372 External BGP Neighbors 375 Checks Before Becoming BGP Neighbors 376 BGP Messages and Neighbor States 378 BGP Message Types 378 Purposefully Resetting BGP Peer Connections 379 Building the BGP Table 380 Injecting Routes/Prefixes into the BGP Table 380 BGP network Command 380 Redistributing from an IGP, Static, or Connected Route 383 Impact of Auto-Summary on Redistributed Routes and the network Command 385 Manual Summaries and the AS_PATH Path Attribute 388 Adding Default Routes to BGP 391 ORIGIN Path Attribute 392 Advertising BGP Routes to Neighbors 393 BGP Update Message 393 Determining the Contents of Updates 394 Example: Impact of the Decision Process and NEXT_HOP on BGP Updates 396 Summary of Rules for Routes Advertised in BGP Updates 402 Building the IP Routing Table 402 Adding eBGP Routes to the IP Routing Table 402 Backdoor Routes 403 Adding iBGP Routes to the IP Routing Table 404 Using Sync and Redistributing Routes 406 Disabling Sync and Using BGP on All Routers in an AS 408 Confederations 409 Configuring Confederations 411 Route Reflectors 414 Foundation Summary 420

xvii

Memory Builders 424 Fill In Key Tables from Memory Definitions 424 Further Reading 425

Chapter 11 BGP Routing Policies

424

427

“Do I Know This Already?” Quiz 427 Foundation Topics 433 Route Filtering and Route Summarization 433 Filtering BGP Updates Based on NLRI 434 Route Map Rules for NLRI Filtering 437 Soft Reconfiguration 438 Comparing BGP Prefix Lists, Distribute Lists, and Route Maps 438 Filtering Subnets of a Summary Using the aggregate-address Command 439 Filtering BGP Updates by Matching the AS_PATH PA 440 The BGP AS_PATH and AS_PATH Segment Types 441 Using Regular Expressions to Match AS_PATH 443 Example: Matching AS_PATHs Using AS_PATH Filters 446 Matching AS_SET and AS_CONFED_SEQ 449 BGP Path Attributes and the BGP Decision Process 452 Generic Terms and Characteristics of BGP PAs 452 The BGP Decision Process 454 Clarifications of the BGP Decision Process 455 Three Final Tiebreaker Steps in the BGP Decision Process 455 Adding Multiple BGP Routes to the IP Routing Table 456 Mnemonics for Memorizing the Decision Process 456 Configuring BGP Policies 458 Background: BGP PAs and Features Used by Routing Policies 458 Step 0: NEXT_HOP Reachable 460 Step 1: Administrative Weight 460 Step 2: Highest Local Preference (LOCAL_PREF) 463 Step 3: Choose Between Locally Injected Routes Based on ORIGIN PA 466 Step 4: Shortest AS_PATH 467 Removing Private ASNs 467 AS_PATH Prepending and Route Aggregation 468 Step 5: Best ORIGIN PA 471 Step 6: Smallest Multi-Exit Discriminator 471 Configuring MED: Single Adjacent AS 473 Configuring MED: Multiple Adjacent Autonomous Systems 474 The Scope of MED 474 Step 7: Prefer Neighbor Type eBGP over iBGP 475 Step 8: Smallest IGP Metric to the NEXT_HOP 475

xviii

The maximum-paths Command and BGP Decision Process Tiebreakers Step 9: Lowest BGP Router ID of Advertising Router (with One Exception) 476 Step 10: Lowest Neighbor ID 476 The BGP maximum-paths Command 476 BGP Communities 478 Matching COMMUNITY with Community Lists 482 Removing COMMUNITY Values 483 Filtering NLRI Using Special COMMUNITY Values 484 Foundation Summary 486 Memory Builders 490 Fill In Key Tables from Memory 490 Definitions 490 Further Reading 490

Part IV QoS

Chapter 12 Classification and Marking

493

“Do I Know This Already?” Quiz 493 Foundation Topics 497 Fields That Can Be Marked for QoS Purposes 497 IP Precedence and DSCP Compared 497 DSCP Settings and Terminology 498 Class Selector PHB and DSCP Values 499 Assured Forwarding PHB and DSCP Values 499 Expedited Forwarding PHB and DSCP Values 500 Non-IP Header Marking Fields 501 Ethernet LAN Class of Service 501 WAN Marking Fields 501 Locations for Marking and Matching 502 Cisco Modular QoS CLI 503 Mechanics of MQC 504 Classification Using Class Maps 505 Using Multiple match Commands 506 Classification Using NBAR 507 Classification and Marking Tools 508 Class-Based Marking (CB Marking) Configuration 508 CB Marking Example 509 CB Marking of CoS and DSCP 513 Network-Based Application Recognition 515 CB Marking Design Choices 516 Marking Using Policers 517 QoS Pre-Classification 518 Policy Routing for Marking 519

475

xix

AutoQoS 519 AutoQoS for VoIP 520 AutoQos VoIP on Switches 520 AutoQoS VoIP on Routers 521 Verifying AutoQoS VoIP 522 AutoQoS for the Enterprise 522 Discovering Traffic for AutoQoS Enterprise 522 Generating the AutoQoS Configuration 523 Verifying AutoQos for the Enterprise 523 Foundation Summary 524 Memory Builders 526 Fill In Key Tables from Memory 526 Definitions 526 Further Reading 527

Chapter 13 Congestion Management and Avoidance

529

“Do I Know This Already?” Quiz 529 Cisco Router Queuing Concepts 533 Software Queues and Hardware Queues 533 Queuing on Interfaces Versus Subinterfaces and Virtual Circuits Comparing Queuing Tools 534 Queuing Tools: CBWFQ and LLQ 535 CBWFQ Basic Features and Configuration 536 Defining and Limiting CBWFQ Bandwidth 538 Low-Latency Queuing 541 Defining and Limiting LLQ Bandwidth 543 LLQ with More Than One Priority Queue 545 Miscellaneous CBWFQ/LLQ Topics 545 Queuing Summary 546 Weighted Random Early Detection 546 How WRED Weights Packets 548 WRED Configuration 549 Modified Deficit Round-Robin 550 LAN Switch Congestion Management and Avoidance 552 Cisco Switch Ingress Queueing 553 Creating a Priority Queue 553 Cisco 3560 Congestion Avoidance 555 Cisco 3560 Switch Egress Queuing 556 Resource Reservation Protocol (RSVP) 559 RSVP Process Overview 560 Configuring RSVP 562 Using RSVP for Voice Calls 563

534

xx

Foundation Summary 565 Memory Builders 565 Fill In Key Tables from Memory Definitions 565 Further Reading 565

565

Chapter 14 Shaping, Policing, and Link Fragmentation

567

“Do I Know This Already?” Quiz 567 Foundation Topics 572 Traffic-Shaping Concepts 572 Shaping Terminology 572 Shaping with an Excess Burst 574 Underlying Mechanics of Shaping 574 Traffic-Shaping Adaptation on Frame Relay Networks 576 Generic Traffic Shaping 576 Class-Based Shaping 578 Tuning Shaping for Voice Using LLQ and a Small Tc 580 Configuring Shaping by Bandwidth Percent 583 CB Shaping to a Peak Rate 584 Adaptive Shaping 584 Frame Relay Traffic Shaping 584 FRTS Configuration Using the traffic-rate Command 586 Setting FRTS Parameters Explicitly 587 FRTS Configuration Using LLQ 588 FRTS Adaptive Shaping 590 FRTS with MQC 590 Policing Concepts and Configuration 590 CB Policing Concepts 591 Single-Rate, Two-Color Policing (One Bucket) 591 Single-Rate, Three-Color Policer (Two Buckets) 592 Two-Rate, Three-Color Policer (Two Buckets) 593 Class-Based Policing Configuration 595 Single-Rate, Three-Color Policing of All Traffic 595 Policing a Subset of the Traffic 596 CB Policing Defaults for Bc and Be 597 Configuring Dual-Rate Policing 597 Multi-Action Policing 597 Policing by Percentage 598 Committed Access Rate 599 QoS Troubleshooting and Commands 601 Troubleshooting Slow Application Response 602 Troubleshooting Voice and Video Problems 603 Other QoS Troubleshooting Tips 604 Approaches to Resolving QoS Issues 605

xxi

Foundation Summary 606 Memory Builders 608 Fill In Key Tables from Memory Definitions 608 Further Reading 609

608

Part V Wide-Area Networks

Chapter 15 Wide-Area Networks

611

“Do I Know This Already?” Quiz 611 Foundation Topics 614 Point-to-Point Protocol 614 PPP Link Control Protocol 615 Basic LCP/PPP Configuration 615 Multilink PPP 617 MLP Link Fragmentation and Interleaving 619 PPP Compression 620 PPP Layer 2 Payload Compression 621 Header Compression 621 Frame Relay Concepts 622 Frame Relay Data Link Connection Identifiers 623 Local Management Interface 624 Frame Relay Headers and Encapsulation 625 Frame Relay Congestion: DE, BECN, and FECN 626 Adaptive Shaping, FECN, and BECN 627 Discard Eligibility Bit 628 Frame Relay Configuration 628 Frame Relay Configuration Basics 629 Frame Relay Payload Compression 632 Frame Relay Fragmentation 634 Frame Relay LFI Using Multilink PPP (MLP) 636 Foundation Summary 638 Memory Builders 641 Fill In Key Tables from Memory 641 Definitions 641

Part VI IP Multicast

Chapter 16 Introduction to IP Multicasting

643

“Do I Know This Already?” Quiz 643 Foundation Topics 646 Why Do You Need Multicasting? 646 Problems with Unicast and Broadcast Methods 647 How Multicasting Provides a Scalable and Manageable Solution

649

xxii

Multicast IP Addresses 652 Multicast Address Range and Structure 652 Well-Known Multicast Addresses 652 Multicast Addresses for Permanent Groups 653 Multicast Addresses for Source-Specific Multicast Applications and Protocols 654 Multicast Addresses for GLOP Addressing 654 Multicast Addresses for Private Multicast Domains 655 Multicast Addresses for Transient Groups 655 Summary of Multicast Address Ranges 655 Mapping IP Multicast Addresses to MAC Addresses 656 Managing Distribution of Multicast Traffic with IGMP 657 Joining a Group 658 Internet Group Management Protocol 659 IGMP Version 2 660 IGMPv2 Host Membership Query Functions 662 IGMPv2 Host Membership Report Functions 663 IGMPv2 Leave Group and Group-Specific Query Messages 666 IGMPv2 Querier 669 IGMPv2 Timers 669 IGMP Version 3 670 LAN Multicast Optimizations 672 Cisco Group Management Protocol 672 IGMP Snooping 678 Router-Port Group Management Protocol 683 Foundation Summary 686 Memory Builders 686 Fill In Key Tables from Memory 687 Definitions 687 Further Reading 687 References in This Chapter 687

Chapter 17 IP Multicast Routing

689

“Do I Know This Already?” Quiz 689 Foundation Topics 693 Multicast Routing Basics 693 Overview of Multicast Routing Protocols 694 Multicast Forwarding Using Dense Mode 694 Reverse Path Forwarding Check 695 Multicast Forwarding Using Sparse Mode 697 Multicast Scoping 699 TTL Scoping 699 Administrative Scoping 700

xxiii

Dense-Mode Routing Protocols 700 Operation of Protocol Independent Multicast Dense Mode 701 Forming PIM Adjacencies Using PIM Hello Messages 701 Source-Based Distribution Trees 702 Prune Message 703 PIM-DM: Reacting to a Failed Link 705 Rules for Pruning 707 Steady-State Operation and the State Refresh Message 709 Graft Message 711 LAN-Specific Issues with PIM-DM and PIM-SM 712 Prune Override 712 Assert Message 713 Designated Router 715 Summary of PIM-DM Messages 715 Distance Vector Multicast Routing Protocol 716 Multicast Open Shortest Path First 716 Sparse-Mode Routing Protocols 717 Operation of Protocol Independent Multicast Sparse Mode 717 Similarities Between PIM-DM and PIM-SM 717 Sources Sending Packets to the Rendezvous Point 718 Joining the Shared Tree 720 Completion of the Source Registration Process 722 Shared Distribution Tree 724 Steady-State Operation by Continuing to Send Joins 725 Examining the RP’s Multicast Routing Table 726 Shortest-Path Tree Switchover 727 Pruning from the Shared Tree 729 Dynamically Finding RPs and Using Redundant RPs 730 Dynamically Finding the RP Using Auto-RP 731 Dynamically Finding the RP Using BSR 735 Anycast RP with MSDP 737 Interdomain Multicast Routing with MSDP 739 Summary: Finding the RP 741 Bidirectional PIM 742 Comparison of PIM-DM and PIM-SM 743 Source-Specific Multicast 744 Foundation Summary 746 Memory Builders 750 Fill In Key Tables from Memory 750 Definitions 751 Further Reading 751

Part VII Security

Chapter 18 Security

753

“Do I Know This Already?” Quiz

753

xxiv

Foundation Topics 757 Router and Switch Device Security 757 Simple Password Protection for the CLI 757 Better Protection of Enable and Username Passwords 758 Using Secure Shell Protocol 759 User Mode and Privileged Mode AAA Authentication 760 Using a Default Set of Authentication Methods 761 Using Multiple Authentication Methods 763 Groups of AAA Servers 764 Overriding the Defaults for Login Security 764 PPP Security 765 Layer 2 Security 766 Switch Security Best Practices for Unused and User Ports 767 Port Security 767 Dynamic ARP Inspection 771 DHCP Snooping 774 IP Source Guard 777 802.1X Authentication Using EAP 777 Storm Control 780 General Layer 2 Security Recommendations 782 Layer 3 Security 783 IP Access Control List Review 784 ACL Rule Summary 785 Wildcard Masks 787 General Layer 3 Security Considerations 788 Smurf Attacks, Directed Broadcasts, and RPF Checks 788 Inappropriate IP Addresses 790 TCP SYN Flood, the Established Bit, and TCP Intercept 790 Classic Cisco IOS Firewall 793 TCP Versus UDP with CBAC 793 Cisco IOS Firewall Protocol Support 794 Cisco IOS Firewall Caveats 794 Cisco IOS Firewall Configuration Steps 795 Cisco IOS Zone-Based Firewall 796 Cisco IOS Intrusion Prevention System 801 Control-Plane Policing 804 Preparing for CoPP Implementation 805 Implementing CoPP 806 Dynamic Multipoint VPN 809 Foundation Summary 811 Memory Builders 814 Fill In Key Tables from Memory 815 Definitions 815 Further Reading 815

xxv

Part VIII MPLS

Chapter 19 Multiprotocol Label Switching

817

“Do I Know This Already?” Quiz 817 Foundation Topics 821 MPLS Unicast IP Forwarding 821 MPLS IP Forwarding: Data Plane 822 CEF Review 822 Overview of MPLS Unicast IP Forwarding 823 MPLS Forwarding Using the FIB and LFIB 825 The MPLS Header and Label 826 The MPLS TTL Field and MPLS TTL Propagation 827 MPLS IP Forwarding: Control Plane 829 MPLS LDP Basics 829 The MPLS Label Information Base Feeding the FIB and LFIB 832 Examples of FIB and LFIB Entries 836 Label Distribution Protocol Reference 838 MPLS VPNs 839 The Problem: Duplicate Customer Address Ranges 840 The Solution: MPLS VPNs 841 MPLS VPN Control Plane 844 Virtual Routing and Forwarding Tables 844 MP-BGP and Route Distinguishers 846 Route Targets 848 Overlapping VPNs 850 MPLS VPN Configuration 851 Configuring the VRF and Associated Interfaces 853 Configuring the IGP Between PE and CE 855 Configuring Redistribution Between PE-CE IGP and MP-BGP 858 Configuring MP-BGP Between PEs 861 MPLS VPN Data Plane 863 Building the (Inner) VPN Label 865 Creating LFIB Entries to Forward Packets to the Egress PE 866 Creating VRF FIB Entries for the Ingress PE 868 Penultimate Hop Popping 869 Other MPLS Applications 870 VRF Lite 872 VRF Lite, Without MPLS 872 VRF Lite with MPLS 875 Foundation Summary 877 Memory Builders 877 Fill In Key Tables from Memory 877 Definitions 877 Further Reading 877

xxvi

Part IX IP Version 6

Chapter 20 IP Version 6

879

“Do I Know This Already?” Quiz 879 Foundation Topics 883 IPv6 Addressing and Address Types 884 IPv6 Address Notation 884 Address Abbreviation Rules 885 IPv6 Address Types 885 Unicast 886 Multicast 889 Anycast 891 The Unspecified Address 892 IPv6 Address Autoconfiguration 892 EUI-64 Address Format 892 Basic IPv6 Functionality Protocols 894 Neighbor Discovery 894 Neighbor Advertisements 896 Neighbor Solicitation 896 Router Advertisement and Router Solicitation 897 Duplicate Address Detection 898 Neighbor Unreachability Detection 899 ICMPv6 899 Unicast Reverse Path Forwarding 900 DNS 901 CDP 901 DHCP 902 Access Lists 903 Traffic Filtering with Access Lists 904 IPv6 Static Routes 904 IPv6 Unicast Routing Protocols 906 OSPFv3 907 Differences Between OSPFv2 and OSPFv3 907 Virtual Links, Address Summarization, and Other OSPFv3 Features OSPFv3 LSA Types 908 OSPFv3 in NBMA Networks 909 Configuring OSPFv3 over Frame Relay 910 Enabling and Configuring OSPFv3 910 Authentication and Encryption 918 EIGRP for IPv6 918 Differences Between EIGRP for IPv4 and for IPv6 918 Unchanged Features 919 Route Filtering 920 Configuring EIGRP for IPv6 920

908

xxvii

Route Redistribution and Filtering 927 IPv6 Route Redistribution 927 Redistribution Example 928 Quality of Service 931 QoS Implementation Strategy 932 Classification, Marking, and Queuing 932 Congestion Avoidance 933 Traffic Shaping and Policing 933 Tunneling Techniques 933 Tunneling Overview 933 Manually Configured Tunnels 935 Automatic IPv4-Compatible Tunnels 936 IPv6 over IPv4 GRE Tunnels 936 Automatic 6to4 Tunnels 937 ISATAP Tunnels 939 NAT-PT 939 IPv6 Multicast 940 Multicast Listener Discovery 940 Explicit Tracking 941 PIM 941 PIM DR Election 941 Source-Specific Multicast 941 PIM BSR 942 Additional PIM Concepts and Options 942 IPv6 Multicast Static Routes 942 Configuring Multicast Routing for IPv6 943 Foundation Summary 944 Memory Builders 946 Fill In Key Tables from Memory 946 Definitions 946 Further Reading 947

Part X Appendixes

Appendix A

Answers to the “Do I Know This Already?” Quizzes

Appendix B

Decimal to Binary Conversion Table

Appendix C

CCIE Exam Updates

Index

986

983

979

949

xxviii

CD-Only

Appendix D

IP Addressing Practice

Appendix E

RIP Version 2

Appendix F

IGMP

Appendix G

Key Tables for CCIE Study

Appendix H

Solutions for Key Tables for CCIE Study

Glossary

xxix

Icons Used in This Book

Communication Server

PC

Terminal

PC with Software

Sun Workstation

File Server

Web Server

Macintosh

Cisco Works Workstation

Branch Office

House, Regular

Headquarters

Printer

Gateway

Laptop

Router

Catalyst Switch

Cisco MDS 9500

Network Cloud

IBM Mainframe

Bridge

Multilayer Switch

Optical Services Router

Label Switch Router

Hub

ATM Switch

ATM router

Route/Switch Processor

Enterprise Fibre Channel disk

Line: Ethernet

Line: Serial

Fibre Channel JBOD

Cluster Controller

Cisco MDS 9500

LAN2LAN Switch

ONS 15540

Line: Switched Serial

xxx

Command Syntax Conventions The conventions used to present command syntax in this book are the same conventions used in the IOS Command Reference. The Command Reference describes these conventions as follows: ■

Boldface indicates commands and keywords that are entered literally as shown. In actual configuration examples and output (not general command syntax), boldface indicates commands that are manually input by the user (such as a show command).

■

Italic indicates arguments for which you supply actual values.

■

Vertical bars (|) separate alternative, mutually exclusive elements.

■

Square brackets ([ ]) indicate an optional element.

■

Braces ({ }) indicate a required choice.

■

Braces within brackets ([{ }]) indicate a required choice within an optional element.

xxxi

Foreword CCIE Routing and Switching Exam Certification Guide, Fourth Edition, is an excellent self-study resource for the CCIE Routing and Switching written exam. Passing this exam is the first step to attaining the valued CCIE Routing and Switching certification and qualifies candidates for the CCIE Routing and Switching lab exam. Gaining certification in Cisco technology is key to the continuing educational development of today’s networking professional. Through certification programs, Cisco validates the skills and expertise required to effectively manage the modern enterprise network. Cisco Press Exam Certification Guides and preparation materials offer exceptional—and flexible—access to the knowledge and information required to stay current in your field of expertise or to gain new skills. Whether used as a supplement to more traditional training or as a primary source of learning, these materials offer users the information and knowledge validation required to gain new understanding and proficiencies. Developed in conjunction with the Cisco certifications and training team, Cisco Press books are the only self-study books authorized by Cisco and offer students a series of exam practice tools and resource materials to help ensure that learners fully grasp the concepts and information presented. Additional authorized Cisco instructor-led courses, e-learning, labs, and simulations are available exclusively from Cisco Learning Solutions Partners worldwide. To learn more, visit http://www.cisco.com/go/training. I hope that you find these materials to be an enriching and useful part of your exam preparation. Erik Ullanderson Manager, Global Certifications Learning@Cisco October 2007

xxxii

Introduction The Cisco Certified Internetwork Expert (CCIE) certification may be the most challenging and prestigious of all networking certifications. It has received numerous awards and certainly has built a reputation as one of the most difficult certifications to earn in all of the technology world. Having a CCIE certification opens doors professionally typically results in higher pay and looks great on a resume. Cisco currently offers several CCIE certifications. This book covers the version 4.0 exam blueprint topics of the written exam for the CCIE Routing and Switching certification. The following list details the currently available CCIE certifications at the time of this book’s publication; check http://www.cisco.com/go/ccie for the latest information. The certifications are listed in the order in which they were made available to the public: ■

CCIE Routing and Switching

■

CCIE Security

■

CCIE Service Provider

■

CCIE Voice

■

CCIE Storage Networking

■

CCIE Wireless

Each of the CCIE certifications requires the candidate to pass both a written exam and a one-day, hands-on lab exam. The written exam is intended to test your knowledge of theory, protocols, and configuration concepts that follow good design practices. The lab exam proves that you can configure and troubleshoot actual gear.

xxxiii

Why Should I Take the CCIE Routing and Switching Written Exam? The first and most obvious reason to take the CCIE Routing and Switching written exam is that it is the first step toward obtaining the CCIE Routing and Switching certification. Also, you cannot schedule a CCIE lab exam until you pass the corresponding written exam. In short, if you want all the professional benefits of a CCIE Routing and Switching certification, you start by passing the written exam. The benefits of getting a CCIE certification are varied, among which are the following: ■

Better pay

■

Career-advancement opportunities

■

Applies to certain minimum requirements for Cisco Silver and Gold Channel Partners, as well as those seeking Master Specialization, making you more valuable to Channel Partners

■

Better movement through the problem-resolution process when calling the Cisco TAC

■

Prestige

■

Credibility for consultants and customer engineers, including the use of the Cisco CCIE logo

The other big reason to take the CCIE Routing and Switching written exam is that it recertifies an individual’s associate-, professional-, and expert-level Cisco certifications. In other words, passing any CCIE written exam recertifies that person’s CCNA, CCNP, CCIP, CCSP, CCDP, and so on. (Recertification requirements do change, so please verify the requirements at http://www.cisco.com/go/certifications.)

xxxiv

CCIE Routing and Switching Written Exam 350-001 The CCIE Routing and Switching written exam, at the time of this writing, consists of a two-hour exam administered at a proctored exam facility affiliated with Pearson VUE (http://www.vue.com/cisco). The exam typically includes approximately 100 multiplechoice questions. No simulation questions are currently part of the written exam. As with most exams, everyone wants to know what is on the exam. Cisco provides general guidance as to topics on the exam in the CCIE Routing and Switching written exam blueprint, the most recent copy of which can be accessed from http://www.cisco.com/ go/ccie. Cisco changes both the CCIE written and lab blueprints over time, but Cisco seldom, if ever, changes the exam numbers. (In contrast, Cisco changes the exam numbers of the associate- and professional-level certifications when it makes major changes to what is covered on those exams.) Instead of changing the exam number when a CCIE exam changes significantly, Cisco publishes a new exam blueprint. Cisco assigns the new blueprint a version number, much like a software version. The CCIE Routing and Switching written exam blueprint 4.0, as of the time of publication, is listed in Table I-1. Table I-1 also lists the chapters that cover each topic. Table I-1

CCIE Routing and Switching Written Exam Blueprint

Topics

Book Chapters

1.00 Implement Layer 2 Technologies 1.10 Implement Spanning Tree Protocol (STP)

3

(a) 802.1d

3

(b) 802.1w

3

(c) 801.1s

3

(d) Loop guard

3

(e) Root guard

3

(f) Bridge protocol data unit (BPDU) guard

3

(g) Storm control

3

(h) Unicast flooding

3

(i) Port roles, failure propagation, and Loop Guard operation

3

1.20 Implement VLAN and VLAN Trunking Protocol (VTP)

2

1.30 Implement trunk and trunk protocols, EtherChannel, and load-balance

2

xxxv

Table I-1

CCIE Routing and Switching Written Exam Blueprint (Continued)

Topics

Book Chapters

1.40 Implement Ethernet technologies

1

(a) Speed and duplex

1

(b) Ethernet, Fast Ethernet, and Gigabit Ethernet

1

(c) PPP over Ethernet (PPPoE)

2

1.50 Implement Switched Port Analyzer (SPAN), Remote Switched Port Analyzer (RSPAN), and flow control

1

1.60 Implement Frame Relay

15

(a) Local Management Interface (LMI)

15

(b) Traffic shaping

15

(c) Full mesh

15

(d) Hub and spoke

15

(e) Discard eligible (DE)

15

1.70 Implement High-Level Data Link Control (HDLC) and PPP

15

2.00 Implement IPv4 2.10 Implement IP version 4 (IPv4) addressing, subnetting, and variable-length subnet masking (VLSM)

4

2.20 Implement IPv4 tunneling and Generic Routing Encapsulation (GRE)

6

2.30 Implement IPv4 RIP version 2 (RIPv2)

E

2.40 Implement IPv4 Open Shortest Path First (OSPF)

8

(a) Standard OSPF areas

8

(b) Stub area

8

(c) Totally stubby area

8

(d) Not-so-stubby-area (NSSA)

8

(e) Totally NSSA

8

(f) Link-state advertisement (LSA) types

8

(g) Adjacency on a point-to-point and on a multi-access network

8

(h) OSPF graceful restart

8

2.50 Implement IPv4 Enhanced Interior Gateway Routing Protocol (EIGRP)

7

(a) Best path

7

(b) Loop-free paths

7

(c) EIGRP operations when alternate loop-free paths are available, and when they are not available

7 continues

xxxvi

Table I-1

CCIE Routing and Switching Written Exam Blueprint (Continued)

Topics

Book Chapters

(d) EIGRP queries

7

(e) Manual summarization and autosummarization

9

(f) EIGRP stubs

7

2.60 Implement IPv4 Border Gateway Protocol (BGP)

10

(a) Next hop

10

(b) Peering

10

(c) Internal Border Gateway Protocol (IBGP) and External Border Gateway Protocol (EBGP)

10, 11

2.70 Implement policy routing

6

2.80 Implement Performance Routing (PfR) and Cisco Optimized Edge Routing (OER)

6

2.90 Implement filtering, route redistribution, summarization, synchronization, attributes, and other advanced

9, 11

3.00 Implement IPv6 3.10 Implement IP version 6 (IPv6) addressing and different addressing types

20

3.20 Implement IPv6 neighbor discovery

20

3.30 Implement basic IPv6 functionality protocols

20

3.40 Implement tunneling techniques

20

3.50 Implement OSPF version 3 (OSPFv3)

20

3.60 Implement EIGRP version 6 (EIGRPv6)

20

3.70 Implement filtering and route redistribution

20

4.00 Implement MPLS Layer 3 VPNs

19

4.10 Implement Multiprotocol Label Switching (MPLS)

19

4.20 Implement Layer 3 virtual private networks (VPNs) on provider edge (PE), provider (P), and customer edge (CE) routers

19

4.30 Implement virtual routing and forwarding (VRF) and Multi-VRF Customer Edge (VRF-Lite)

19

5.00 Implement IP Multicast 5.10 Implement Protocol Independent Multicast (PIM) sparse mode

16, 17

5.20 Implement Multicast Source Discovery Protocol (MSDP)

17

5.30 Implement interdomain multicast routing

17

5.40 Implement PIM Auto-Rendezvous Point (Auto-RP), unicast rendezvous point (RP), and bootstrap router (BSR)

17

xxxvii

Table I-1

CCIE Routing and Switching Written Exam Blueprint (Continued)

Topics

Book Chapters

5.50 Implement multicast tools, features, and source-specific multicast

17

5.60 Implement IPv6 multicast, PIM, and related multicast protocols, such as Multicast Listener Discovery (MLD)

17

6.00 Implement Network Security 6.01 Implement access lists

18

6.02 Implement Zone Based Firewall

18

6.03 Implement Unicast Reverse Path Forwarding (uRPF)

18

6.04 Implement IP Source Guard

18

6.05 Implement authentication, authorization, and accounting (AAA) (configuring the AAA server is not required, only the client side (IOS) is configured)

18

6.06 Implement Control Plane Policing (CoPP)

18

6.07 Implement Cisco IOS Firewall

18

6.08 Implement Cisco IOS Intrusion Prevention System (IPS)

18

6.09 Implement Secure Shell (SSH)

18

6.10 Implement 802.1x

18

6.11 Implement NAT

18

6.12 Implement routing protocol authentication

18

6.13 Implement device access control

18

6.14 Implement security features

18

7.00 Implement Network Services 7.10 Implement Hot Standby Router Protocol (HSRP)

5

7.20 Implement Gateway Load Balancing Protocol (GLBP)

5

7.30 Implement Virtual Router Redundancy Protocol (VRRP)

5

7.40 Implement Network Time Protocol (NTP)

5

7.50 Implement DHCP

5

7.60 Implement Web Cache Communication Protocol (WCCP)

5

8.00 Implement Quality of Service (QoS) 8.10 Implement Modular QoS CLI (MQC)

12

(a) Network-Based Application Recognition (NBAR)

12

(b) Class-based weighted fair queuing (CBWFQ), modified deficit round robin (MDRR), and low latency queuing (LLQ)

13

(c) Classification

12 continues

xxxviii

Table I-1

CCIE Routing and Switching Written Exam Blueprint (Continued)

Topics

Book Chapters

(d) Policing

14

(e) Shaping

14

(f) Marking

12

(g) Weighted random early detection (WRED) and random early detection (RED)

13

(h) Compression

15

8.20 Implement Layer 2 QoS: weighted round robin (WRR), shaped round robin (SRR), and policies

13

8.30 Implement link fragmentation and interleaving (LFI) for Frame Relay

15

8.40 Implement generic traffic shaping

14

8.50 Implement Resource Reservation Protocol (RSVP)

13

8.60 Implement Cisco AutoQoS

12

9.00 Troubleshoot a Network 9.10 Troubleshoot complex Layer 2 network issues

3

9.20 Troubleshoot complex Layer 3 network issues

9

9.30 Troubleshoot a network in response to application problems

14

9.40 Troubleshoot network services

6

9.50 Troubleshoot network security

18

10.00 Optimize the Network 10.01 Implement syslog and local logging

5

10.02 Implement IP Service Level Agreement SLA

5

10.03 Implement NetFlow

5

10.04 Implement SPAN, RSPAN, and router IP traffic export (RITE)

5

10.05 Implement Simple Network Management Protocol (SNMP)

5

10.06 Implement Cisco IOS Embedded Event Manager (EEM)

5

10.07 Implement Remote Monitoring (RMON)

5

10.08 Implement FTP

5

10.09 Implement TFTP

5

10.10 Implement TFTP server on router

5

10.11 Implement Secure Copy Protocol (SCP)

5

10.12 Implement HTTP and HTTPS

5

10.13 Implement Telnet

5

xxxix

Table I-1

CCIE Routing and Switching Written Exam Blueprint (Continued) Book Chapters

Topics

11.00 Evaluate proposed changes to a Network 11.01 Evaluate interoperability of proposed technologies against deployed technologies

N/A

(a) Changes to routing protocol parameters

N/A

(b) Migrate parts of a network to IPv6

N/A

(c) Routing Protocol migration

N/A

(d) Adding multicast support

N/A

(e) Migrate spanning tree protocol

N/A

(f) Evaluate impact of new traffic on existing QoS design

N/A

11.02 Determine operational impact of proposed changes to an existing network

N/A

(a) Downtime of network or portions of network

N/A

(b) Performance degradation

N/A

(c) Introducing security breaches

N/A

11.03 Suggest Alternative solutions when incompatible changes are proposed to an existing network

N/A

(a) Hardware/Software upgrades

N/A

(b) Topology shifts

N/A

(c) Reconfigurations

N/A

Version 4.0 of the blueprint provides more detail than the earlier versions of the blueprint. It is also helpful to know what topics Cisco has removed from earlier blueprints, because it is also useful to know what not to study as well as what to study. The more significant topics removed from the last few versions of the CCIE R/S Written blueprints include the following: ■

Version 2.0 (2005)—Cisco announced the removal of ISDN/DDR, IS-IS, ATM, and SONET; they also added wireless LANs

■

Version 3.0 (2007)—The Version 3.0 blueprint showed the removal of wireless LANs, and added IPv6 and MPLS concepts.

■

Version 4.0 (2009)—The Version 4.0 blueprint shows that no significant topics were removed.

The Version 4.0 blueprint adds many new topics compared to the Version 3.0 blueprint. The blueprint mentions around 20 new small topics. In addition, the blueprint wording has been changed to be more aligned with the other Cisco certifications, with many of the topics listing the word configuration. Notably, MPLS configuration has been added since

xl

Version 3.0, with several of the small topics, ranging in one to three pages of coverage in the book, also now including some configuration discussion. The Version 4.0 blueprint also now includes five troubleshooting topics, as listed in section 9.0 of the blueprint, and paraphrased as follows: ■

LANs

■

IP routing

■

Application performance (QoS)

■

Network services

■

Security

The existence of specific topics for troubleshooting may be a bit confusing at first, because the CCIE lab also now contains a specific troubleshooting component. However, the prior versions of the CCIE written exam already included questions asked in the context of a broken network or misconfigured device. These new blueprint items simply formalize the idea that you should not only understand proper configuration, but be able to predict what will happen when problems occur. Finally, the other big change between the Version 3.0 and Version 4.0 blueprint relates to section 11.0 of the blueprint. This new section might be better termed “Dealing with issues that arise in real life when networks change.” Section 11.0, actually titled “Evaluate Proposed Changes to a Network,” diverges from the usual convention of a list of specific technologies. Instead, section 11.0 lists topics about how engineers do their jobs. Specifically, these topics relate to issues that arise when implementing network technologies in an existing network—topics that can be well learned by doing a network engineering job, and questions that can be answered by applying the vast amount of information covered through the whole book. From one perspective, the whole book already covers the topics in this section, but there is no specific section of the printed book that addresses these topics. To give you practice on these topics, and pull the topics together, Edition 4 of the CCIE Routing and Switching Exam Certification Guide includes a large set of CD questions that mirror the types of questions expected for part 11 of the Version 4.0 blueprint. By their very nature, these topics require the application of the knowledge listed throughout the book. This special section of questions provides a means to learn and practice these skills with a proportionally larger set of questions added specifically for this purpose. These questions will be available to you in the practice test engine database, whether you take full exams or choose questions by category.

xli

About the CCIE Routing and Switching Official Exam Certification Guide, Fourth Edition This section provides a brief insight into the contents of the book, the major goals, and some of the book features that you will encounter when using this book.

Book Organization This book contains nine major parts. The book places the longer and the more long-lived topics earlier in the book. For example, the most familiar topics, LAN switching and IPv4 routing, occupy the first three parts, and consume more than 400 pages of the book. QoS, which has been a part of the blueprint for a long times, follows as part IV. Beyond the chapters in the nine major parts of the book, you will find several useful appendixes gathered in Part X. Following is a description of each part’s coverage: ■

Part I, “LAN Switching” (Chapters 1–3) This part focuses on LAN Layer 2 features, specifically Ethernet (Chapter 1), VLANs and trunking (Chapter 2), and Spanning Tree Protocol (Chapter 3).

■

Part II, “IP” (Chapters 4–5) This part is titled “IP” to match the blueprint, but it might be better titled “TCP/IP” because it covers details across the spectrum of the TCP/IP protocol stack. It includes IP addressing (Chapter 4) and IP services such as DHCP and ARP (Chapter 5).

■

Part III, “IP Routing” (Chapters 6–11) This part covers some of the more important topics on the exam and is easily the largest part of the book. It covers Layer 3 forwarding concepts (Chapter 6), followed by two routing protocol chapters, one each about EIGRP and OSPF (Chapters 7 and 8, respectively). (Note that while RIP Version 2 is listed in the blueprint, its role is waning; therefore, that material exists in this book as CD-only Appendix E.) Following that, Chapter 9 covers route redistribution between IGPs. At the end, Chapter 10 hits the details of BGP, with Chapter 11 looking at BGP path attributes and how to influence BGP’s choice of best path.

■

Part IV, “QoS” (Chapters 12–14) This part covers the more popular QoS tools, including some MQC-based tools, as well as several older tools, particularly FRTS. The chapters include coverage of classification and marking (Chapter 12), queuing and congestion avoidance (Chapter 13), plus shaping, policing, and link efficiency (Chapter 14).

xlii

■

Part V, “Wide-Area Networks” (Chapter 15) The WAN coverage has been shrinking over the last few revisions to the CCIE R&S written exam. Chapter 15 includes some brief coverage of PPP and Frame Relay. Note that the previous version (V3.0) and current version (V4.0) of the blueprint includes another WAN topic, MPLS, which is covered in Part VIII, Chapter 19.

■

Part VI, “IP Multicast” (Chapters 16–17) Chapter 16 covers multicast on LANs, including IGMP and how hosts join multicast groups. Chapter 17 covers multicast WAN topics.

■

Part VII, “Security” (Chapter 18) Given the CCIE tracks for both Security and Voice, Cisco has a small dilemma regarding whether to cover those topics on CCIE Routing and Switching, and if so, in how much detail. This part covers a variety of security topics appropriate for CCIE Routing and Switching, in a single chapter. This chapter focuses on switch and router security.

■

Part VIII, “MPLS” (Chapter 19) As mentioned in the WAN section, the CCIE R&S exam’s coverage of MPLS has been growing over the last two versions of the blueprint. This chapter focuses on enterpriserelated topics such as core MPLS concepts and MPLS VPNs, including basic configuration.

■

Part IX, “IP Version 6” (Chapter 20) Chapter 20 examines a wide variety of IPv6 topics, including addressing, routing protocols, redistribution, and coexistence.

■

Part X, “Appendixes” Appendix A, “Answers to the ‘Do I Know This Already?’ Quizzes” This appendix lists answers and explanations for the questions at the beginning of each chapter. Appendix B, “Decimal to Binary Conversion Table” This appendix lists the decimal values 0 through 255, with their binary equivalents. Appendix C, “CCIE Routing and Switching Exam Updates: Version 1.0” As of the first printing of the book, this appendix contains only a few words that reference the web page for this book at http://www.ciscopress.com/title/9781587059803. As the blueprint

xliii

evolves over time, the authors will post new materials at the website. Any future printings of the book will include the latest newly added materials in printed form inside Appendix C. If Cisco releases a major exam update, changes to the book will be available only in a new edition of the book and not on this site. NOTE Appendixes D through H and the Glossary are in printable, PDF format on

the CD.

(CD-only) Appendix D, “IP Addressing Practice” This appendix lists several practice problems for IP subnetting and finding summary routes. The explanations to the answers use the shortcuts described in the book. (CD-only) Appendix E, “RIP Version 2” This appendix lists a copy of the RIP Version 2 chapter from the previous edition of this book. (CD-only) Appendix F, “IGMP” This short appendix contains background information on Internet Group Management Protocol (IGMP) that was in the previous edition’s first multicast chapter. It is included in case the background information might be useful to some readers. (CD-only) Appendix G, “Key Tables for CCIE Study” This appendix lists the most important tables from the core chapters of the book. The tables have much of the content removed so that you can use them as an exercise. You can print the PDF and then fill in the table from memory, checking your answers against the completed tables in Appendix H. (CD-only) Glossary The Glossary contains the key terms listed in the book.

Book Features The core chapters of this book have several features that help you make the best use of your time: ■

“Do I Know This Already?” Quizzes—Each chapter begins with a quiz that helps you to determine the amount of time you need to spend studying that chapter. If you score yourself strictly, and you miss only one question, you may want to skip the core

xliv

of the chapter and move on to the “Foundation Summary” section at the end of the chapter, which lets you review facts and spend time on other topics. If you miss more than one, you may want to spend some time reading the chapter or at least reading sections that cover topics about which you know you are weaker. ■

Foundation Topics—These are the core sections of each chapter. They explain the protocols, concepts, and configuration for the topics in that chapter.

■

Foundation Summary—The “Foundation Summary” section of this book departs from the typical features of the “Foundation Summary” section of other Cisco Press Exam Certification Guides. This section does not repeat any details from the “Foundation Topics” section; instead, it simply summarizes and lists facts related to the chapter but for which a longer or more detailed explanation is not warranted.

■

Key topics—Throughout the “Foundation Topics” section, a Key Topic icon has been placed beside the most important areas for review. After reading a chapter, when doing your final preparation for the exam, take the time to flip through the chapters, looking for the Key Topic icons, and review those paragraphs, tables, figures, and lists.

■

Fill In Key Tables from Memory—The more important tables from the chapters have been copied to PDF files available on the CD as Appendix G. The tables have most of the information removed. After printing these mostly empty tables, you can use them to improve your memory of the facts in the table by trying to fill them out. This tool should be useful for memorizing key facts. That same CD-only appendix contains the completed tables so you can check your work.

■

CD-based practice exam—The companion CD contains multiple-choice questions and a testing engine. The CD includes 200 questions unique to the CD. As part of your final preparation, you should practice with these questions to help you get used to the exam-taking process, as well as help refine and prove your knowledge of the exam topics.

■

Special question section for the “Implement Proposed Changes to a Network” section of the Blueprint—To provide practice and perspectives on these exam topics, a special section of questions has been developed to help you both prepare for these new types of questions.

xlv

■

Key terms and Glossary—The more important terms mentioned in each chapter are listed at the end of each chapter under the heading “Definitions.” The Glossary, found on the CD that comes with this book, lists all the terms from the chapters. When studying each chapter, you should review the key terms, and for those terms about which you are unsure of the definition, you can review the short definitions from the Glossary.

■

Further Reading—Most chapters include a suggested set of books and websites for additional study on the same topics covered in that chapter. Often, these references will be useful tools for preparation for the CCIE Routing and Switching lab exam.

Blueprint topics covered in this chapter: This chapter covers the following subtopics from the Cisco CCIE Routing and Switching written exam blueprint. Refer to the full blueprint in Table I-1 in the Introduction for more details on the topics covered in each chapter and their context within the blueprint. ■

Ethernet

■

Speed

■

Duplex

■

Fast Ethernet

■

Gigabit Ethernet

■

SPAN and RSPAN

CHAPTER

1

Ethernet Basics It’s no surprise that the concepts, protocols, and commands related to Ethernet are a key part of the CCIE Routing and Switching written exam. Almost all campus networks today are built using Ethernet technology. Also, Ethernet technology is moving into the WAN with the emergence of metro Ethernet. Even in an IT world, where technology changes rapidly, you can expect that ten years from now, Ethernet will still be an important part of the CCIE Routing and Switching written and lab exams. For this chapter, if I had to venture a guess, probably 100 percent of you reading this book know a fair amount about Ethernet basics already. I must admit, I was tempted to leave it out. However, I would also venture a guess that at least some of you have forgotten a few facts about Ethernet. So you can read the whole chapter if your Ethernet recollections are a bit fuzzy— or you could just hit the highlights. For exam preparation, it is typically useful to use all the refresher tools: take the “Do I Know This Already?” quiz, complete the definitions of the terms listed at the end of the chapter, print and complete the tables in Appendix G, “Key Tables for CCIE Study,” and certainly answer all the CD-ROM questions concerning Ethernet.

“Do I Know This Already?” Quiz Table 1-1 outlines the major headings in this chapter and the corresponding “Do I Know This Already?” quiz questions. Table 1-1

“Do I Know This Already?” Foundation Topics Section-to-Question Mapping Foundation Topics Section

Questions Covered in This Section

Ethernet Layer 1: Wiring, Speed, and Duplex

1–5

Ethernet Layer 2: Framing and Addressing

6–7

Switching and Bridging Logic

8

SPAN and RSPAN

9–10

Total Score

Score

4

Chapter 1: Ethernet Basics

In order to best use this pre-chapter assessment, remember to score yourself strictly. You can find the answers in Appendix A, “Answers to the ‘Do I Know This Already?’ Quizzes.” 1.

2.

3.

4.

Which of the following denotes the correct usage of pins on the RJ-45 connectors at the opposite ends of an Ethernet cross-over cable? a.

1 to 1

b.

1 to 2

c.

1 to 3

d.

6 to 1

e.

6 to 2

f.

6 to 3

Which of the following denotes the correct usage of pins on the RJ-45 connectors at the opposite ends of an Ethernet straight-through cable? a.

1 to 1

b.

1 to 2

c.

1 to 3

d.

6 to 1

e.

6 to 2

f.

6 to 3

Which of the following commands must be configured on a Cisco IOS switch interface to disable Ethernet auto-negotiation? a.

no auto-negotiate

b.

no auto

c.

Both speed and duplex

d.

duplex

e.

speed

Consider an Ethernet cross-over cable between two 10/100 ports on Cisco switches. One switch has been configured for 100-Mbps full duplex. Which of the following is true about the other switch? a.

It will use a speed of 10 Mbps.

b.

It will use a speed of 100 Mbps.

c.

It will use a duplex setting of half duplex.

d.

It will use a duplex setting of full duplex.

“Do I Know This Already?” Quiz

5.

6.

7.

8.

9.

5

Consider an Ethernet cross-over cable between two 10/100/1000 ports on Cisco switches. One switch has been configured for half duplex, and the other for full duplex. The ports successfully negotiate a speed of 1 Gbps. Which of the following could occur as a result of the duplex mismatch? a.

No frames can be received by the half-duplex switch without it believing an FCS error has occurred.

b.

CDP would detect the mismatch and change the full-duplex switch to half duplex.

c.

CDP would detect the mismatch and issue a log message to that effect.

d.

The half-duplex switch will erroneously believe collisions have occurred.

Which of the following Ethernet header type fields is a 2-byte field? a.

DSAP

b.

Type (in SNAP header)

c.

Type (in Ethernet V2 header)

d.

LLC Control

Which of the following standards defines a Fast Ethernet standard? a.

IEEE 802.1Q

b.

IEEE 802.3U

c.

IEEE 802.1X

d.

IEEE 802.3Z

e.

IEEE 802.3AB

f.

IEEE 802.1AD

Suppose a brand-new Cisco IOS–based switch has just been taken out of the box and cabled to several devices. One of the devices sends a frame. For which of the following destinations would a switch flood the frames out all ports (except the port upon which the frame was received)? a.

Broadcasts

b.

Unknown unicasts

c.

Known unicasts

d.

Multicasts

Which of the following configuration issues will keep a SPAN session from becoming active? a.

Misconfigured destination port

b.

Destination port configured as a trunk

c.

Destination port shutdown

d.

Source port configured as a trunk

6

Chapter 1: Ethernet Basics

10.

Which of the following are rules for SPAN configuration? a.

SPAN source and destination ports must be configured for the same speed and duplex.

b.

If the SPAN source port is configured for 100 Mbps, the destination port must be configured for 100 Mbps or more.

c.

In a SPAN session, sources must consist of either physical interfaces or VLANs, but not a mix of these.

d.

Remote SPAN VLANs must be in the range of VLAN 1–66.

e.

Only three SPAN sessions may be configured on one switch.

Ethernet Layer 1: Wiring, Speed, and Duplex

7

Foundation Topics

Ethernet Layer 1: Wiring, Speed, and Duplex Before making an Ethernet LAN functional, end-user devices, routers, and switches must be cabled correctly. To run with fewer transmission errors at higher speeds, and to support longer cable distances, variations of copper and optical cabling can be used. The different Ethernet specifications, cable types, and cable lengths per the various specifications are important for the exam, and are listed in the “Foundation Summary” section.

RJ-45 Pinouts and Category 5 Wiring You should know the details of cross-over and straight-through Category 5 (Cat 5) or Cat 5e cabling for most any networking job. The EIA/TIA defines the cabling specifications for Ethernet LANs (http://www.eia.org and http://www.tiaonline.org), including the pinouts for the RJ-45 connects, as shown in Figure 1-1. Figure 1-1

RJ-45 Pinouts with Four-Pair UTP Cabling

The most popular Ethernet standards (10BASE-T and 100BASE-TX) each use two twisted pairs (specifically pairs 2 and 3 shown in Figure 1-1), with one pair used for transmission in each direction. Depending on which pair a device uses to transmit and receive, either a straight-through or cross-over cable is required. Table 1-2 summarizes how the cabling and pinouts work. Table 1-2

Ethernet Cabling Types Type of Cable

Pinouts

Key Pins Connected

Straight-through

T568A (both ends) or T568B (both ends)

1 – 1; 2 – 2; 3 – 3; 6 – 6

Cross-over

T568A on one end, T568B on the other

1 – 3; 2 – 6; 3 – 1; 6 – 2

8

Chapter 1: Ethernet Basics

Many Ethernet standards use two twisted pairs, with one pair being used for transmission in each direction. For instance, a PC network interface card (NIC) transmits on pair 1,2 and receives on pair 3,6; switch ports do the opposite. So, a straight-through cable works well, connecting pair 1,2 on the PC (PC transmit pair) to the switch port’s pair 1,2, on which the switch receives. When the two devices on the ends of the cable both transmit using the same pins, a cross-over cable is required. For instance, if two connected switches send using the pair at pins 3,6 and receive on pins 1,2, then the cable needs to connect the pair at 3,6 on one end to pins 1,2 at the other end, and vice versa. NOTE Cross-over cables can also be used between a pair of PCs, swapping the transmit pair on one end (1,2) with the receive pins at the other end (3,6).

Cisco also supports a switch feature that lets the switch figure out if the wrong cable is installed: Auto-MDIX (automatic medium-dependent interface crossover) detects the wrong cable and causes the switch to swap the pair it uses for transmitting and receiving, which solves the cabling problem. (As of publication, this feature is not supported on all Cisco switch models.)

Auto-negotiation, Speed, and Duplex By default, each Cisco switch port uses Ethernet auto-negotiation to determine the speed and duplex setting (half or full). The switches can also set their duplex setting with the duplex interface subcommand, and their speed with—you guessed it—the speed interface subcommand. Switches can dynamically detect the speed setting on a particular Ethernet segment by using a few different methods. Cisco switches (and many other devices) can sense the speed using the Fast Link Pulses (FLP) of the auto-negotiation process. However, if auto-negotiation is disabled on either end of the cable, the switch detects the speed anyway based on the incoming electrical signal. You can force a speed mismatch by statically configuring different speeds on either end of the cable, causing the link to no longer function. Switches detect duplex settings through auto-negotiation only. If both ends have autonegotiation enabled, the duplex is negotiated. However, if either device on the cable disables auto-negotiation, the devices without a configured duplex setting must assume a default. Cisco switches use a default duplex setting of half duplex (HDX) (for 10-Mbps and 100-Mbps interfaces) or full duplex (FDX) (for 1000-Mbps interfaces). To disable auto-negotiation on a Cisco switch port, you simply need to statically configure the speed and the duplex settings. Ethernet devices can use FDX only when collisions cannot occur on the attached cable; a collision-free link can be guaranteed only when a shared hub is not in use. The next few topics review how Ethernet deals with collisions when they do occur, as well as what is different with Ethernet logic in cases where collisions cannot occur and FDX is allowed.

Ethernet Layer 1: Wiring, Speed, and Duplex

9

CSMA/CD The original Ethernet specifications expected collisions to occur on the LAN. The media was shared, creating a literal electrical bus. Any electrical signal induced onto the wire could collide with a signal induced by another device. When two or more Ethernet frames overlap on the transmission medium at the same instant in time, a collision occurs; the collision results in bit errors and lost frames. The original Ethernet specifications defined the Carrier Sense Multiple Access with Collision Detection (CSMA/CD) algorithm to deal with the inevitable collisions. CSMA/CD minimizes the number of collisions, but when they occur, CSMA/CD defines how the sending stations can recognize the collisions and retransmit the frame. The following list outlines the steps in the CSMA/CD process: 1.

A device with a frame to send listens until the Ethernet is not busy (in other words, the device cannot sense a carrier signal on the Ethernet segment).

2.

When the Ethernet is not busy, the sender begins sending the frame.

3.

The sender listens to make sure that no collision occurred.

4.

If there was a collision, all stations that sent a frame send a jamming signal to ensure that all stations recognize the collision.

5.

After the jamming is complete, each sender of one of the original collided frames randomizes a timer and waits that long before resending. (Other stations that did not create the collision do not have to wait to send.)

6.

After all timers expire, the original senders can begin again with Step 1.

Collision Domains and Switch Buffering A collision domain is a set of devices that can send frames that collide with frames sent by another device in that same set of devices. Before the advent of LAN switches, Ethernets were either physically shared (10BASE2 and 10BASE5) or shared by virtue of shared hubs and their Layer 1 “repeat out all other ports” logic. Ethernet switches greatly reduce the number of possible collisions, both through frame buffering and through their more complete Layer 2 logic. By definition of the term, Ethernet hubs: ■

Operate solely at Ethernet Layer 1

■

Repeat (regenerate) electrical signals to improve cabling distances

■

Forward signals received on a port out all other ports (no buffering)

As a result of a hub’s logic, a hub creates a single collision domain. Switches, however, create a different collision domain per switch port, as shown in Figure 1-2.

10

Chapter 1: Ethernet Basics

Figure 1-2

Collision Domains with Hubs and Switches 1 Collision Domain 10BASE-T, using Shared hub

Multiple Collision Domain 10BASE-T, using Switch

Archie

Larry

Hub1

Archie

Larry Bob

Solid Lines Represent Twisted Pair Cabling

SW1 Bob

Switches have the same cabling and signal regeneration benefits as hubs, but switches do a lot more—including sometimes reducing or even eliminating collisions by buffering frames. When switches receive multiple frames on different switch ports, they store the frames in memory buffers to prevent collisions. For instance, imagine that a switch receives three frames at the same time, entering three different ports, and they all must exit the same switch port. The switch simply stores two of the frames in memory, forwarding the frames sequentially. As a result, in Figure 1-2, the switch prevents any frame sent by Larry from colliding with a frame sent by Archie or Bob—which by definition puts each of the PCs attached to the switch in Figure 1-2 in different collision domains. When a switch port connects via cable to a single other non-hub device—for instance, like the three PCs in Figure 1-2—no collisions can possibly occur. The only devices that could create a collision are the switch port and the one connected device—and they each have a separate twisted pair on which to transmit. Because collisions cannot occur, such segments can use full-duplex logic. When a switch port connects to a hub, it needs to operate in HDX mode, because collisions might occur due to the logic used by the hub. NOTE NICs operating in HDX mode use loopback circuitry when transmitting a frame. This circuitry loops the transmitted frame back to the receive side of the NIC, so that when the NIC receives a frame over the cable, the combined looped-back signal and received signal allows the NIC to notice that a collision has occurred.

Ethernet Layer 1: Wiring, Speed, and Duplex

11

Basic Switch Port Configuration The three key configuration elements on a Cisco switch port are auto-negotiation, speed, and duplex. Cisco switches use auto-negotiation by default; it is then disabled if both the speed and duplex are manually configured. You can set the speed using the speed {auto | 10 | 100 | 1000} interface subcommand, assuming the interface supports multiple speeds. You configure the duplex setting using the duplex {auto | half | full} interface subcommand. Example 1-1 shows the manual configuration of the speed and duplex on the link between Switch1 and Switch4 from Figure 1-3, and the results of having mismatched duplex settings. (The book refers to specific switch commands used on IOS-based switches, referred to as “Catalyst IOS” by the Cisco CCIE blueprint.) Figure 1-3

Simple Switched Network with Trunk 0200.3333.3333

0/3

SW1

R3

0/13 000a.b7dc.b78d 000f.2343.87cd 0/13 0200.4444.4444

0/4

R4

SW4 0/6

0010.a49b.6111

PC1

Example 1-1

Manual Setting for Duplex and Speed, with Mismatched Duplex

switch1# show interface fa 0/13 FastEthernet0/13 is up, line protocol is up Hardware is Fast Ethernet, address is 000a.b7dc.b78d (bia 000a.b7dc.b78d) MTU 1500 bytes, BW 100000 Kbit, DLY 100 usec, reliability 255/255, txload 1/255, rxload 1/255 Encapsulation ARPA, loopback not set Keepalive set (10 sec) Full-duplex, 100Mb/s ! remaining lines omitted for brevity ! Below, Switch1’s interface connecting to Switch4 is configured for 100 Mbps, ! HDX. Note that IOS rejects the first duplex command; you cannot set duplex until ! the speed is manually configured. switch1# conf t Enter configuration commands, one per line.

End with CNTL/Z.

switch1(config)# int fa 0/13 switch1(config-if)# duplex half Duplex will not be set until speed is set to non-auto value switch1(config-if)# speed 100

continues

12

Chapter 1: Ethernet Basics

Example 1-1

Manual Setting for Duplex and Speed, with Mismatched Duplex (Continued)

05:08:41: %LINEPROTO-5-UPDOWN: Line protocol on Interface FastEthernet0/13, changed state to down 05:08:46: %LINEPROTO-5-UPDOWN: Line protocol on Interface FastEthernet0/13, changed state to up switch1(config-if)# duplex half !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! ! NOT SHOWN: Configuration for 100/half on Switch4’s int fa 0/13. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! ! Now with both switches manually configured for speed and duplex, neither will be ! using Ethernet auto-negotiation. As a result, below the duplex setting on Switch1 ! can be changed to FDX with Switch4 remaining configured to use HDX. switch1# conf t Enter configuration commands, one per line.

End with CNTL/Z.

switch1(config)# int fa 0/13 switch1(config-if)# duplex full 05:13:03: %LINEPROTO-5-UPDOWN: Line protocol on Interface FastEthernet0/13, changed state to down 05:13:08: %LINEPROTO-5-UPDOWN: Line protocol on Interface FastEthernet0/13, changed state to up ^Z switch1(config-if)#^ switch1# sh int fa 0/13 FastEthernet0/13 is up, line protocol is up ! Lines omitted for brevity Full-duplex, 100Mb/s ! remaining lines omitted for brevity ! Below, Switch4 is shown to be HDX. Note ! the collisions counters at the end of the show interface command. switch4# sh int fa 0/13 FastEthernet0/13 is up, line protocol is up (connected) Hardware is Fast Ethernet, address is 000f.2343.87cd (bia 000f.2343.87cd) MTU 1500 bytes, BW 100000 Kbit, DLY 1000 usec, reliability 255/255, txload 1/255, rxload 1/255 Encapsulation ARPA, loopback not set Keepalive set (10 sec) Half-duplex, 100Mb/s ! Lines omitted for brevity 5 minute output rate 583000 bits/sec, 117 packets/sec 25654 packets input, 19935915 bytes, 0 no buffer Received 173 broadcasts (0 multicast) 0 runts, 0 giants, 0 throttles 0 input errors, 0 CRC, 0 frame, 0 overrun, 0 ignored 0 watchdog, 173 multicast, 0 pause input 0 input packets with dribble condition detected 26151 packets output, 19608901 bytes, 0 underruns 54 output errors, 5 collisions, 0 interface resets 0 babbles, 54 late collision, 59 deferred 0 lost carrier, 0 no carrier, 0 PAUSE output 0 output buffer failures, 0 output buffers swapped out

Ethernet Layer 2: Framing and Addressing

Example 1-1

13

Manual Setting for Duplex and Speed, with Mismatched Duplex (Continued)

02:40:49: %CDP-4-DUPLEX_MISMATCH: duplex mismatch discovered on FastEthernet0/13 (not full duplex), with Switch1 FastEthernet0/13 (full duplex). ! Above, CDP messages have been exchanged over the link between switches. CDP ! exchanges information about Duplex on the link, and can notice (but not fix) ! the mismatch.

The statistics on switch4 near the end of the example show collisions (detected in the time during which the first 64 bytes were being transmitted) and late collisions (after the first 64 bytes were transmitted). In an Ethernet that follows cabling length restrictions, collisions should be detected while the first 64 bytes are being transmitted. In this case, Switch1 is using FDX logic, meaning it sends frames anytime—including when Switch4 is sending frames. As a result, Switch4 receives frames anytime, and if sending at the time, it believes a collision has occurred. Switch4 has deferred 59 frames, meaning that it chose to wait before sending frames because it was currently receiving a frame. Also, the retransmission of the frames that Switch4 thought were destroyed due to a collision, but may not have been, causes duplicate frames to be received, occasionally causing application connections to fail and routers to lose neighbor relationships.

Ethernet Layer 2: Framing and Addressing In this book, as in many Cisco courses and documents, the word frame refers to the bits and bytes that include the Layer 2 header and trailer, along with the data encapsulated by that header and trailer. The term packet is most often used to describe the Layer 3 header and data, without a Layer 2 header or trailer. Ethernet’s Layer 2 specifications relate to the creation, forwarding, reception, and interpretation of Ethernet frames. The original Ethernet specifications were owned by the combination of Digital Equipment Corp., Intel, and Xerox—hence the name “Ethernet (DIX).” Later, in the early 1980s, the IEEE standardized Ethernet, defining parts (Layer 1 and some of Layer 2) in the 802.3 Media Access Control (MAC) standard, and other parts of Layer 2 in the 802.2 Logical Link Control (LLC) standard. Later, the IEEE realized that the 1-byte DSAP field in the 802.2 LLC header was too small. As a result, the IEEE introduced a new frame format with a Sub-Network Access Protocol (SNAP) header after the 802.2 header, as shown in the third style of header in Figure 1-4. Finally, in 1997, the IEEE added the original DIX V2 framing to the 802.3 standard as well as shown in the top frame in Figure 1-40. Table 1-3 lists the header fields, along with a brief explanation. The more important fields are explained in more detail after the table.

14

Chapter 1: Ethernet Basics

Figure 1-4

Ethernet Framing Options Ethernet (DIX) and Revised (1997) IEEE 802.3 8 6 6 2 Variable 4 Preamble

Dest. Source Address Address

Original IEEE Ethernet (802.3) 7 1 6 6

Type/ Data FCS Length

2

1 1 1-2 Variable 4 D S Dest. Source S S Preamble SD Length A A Control Data FCS address address P P 802.3

802.2

IEEE 802.3 with SNAP Header 7 1 6 6

2

1 D Dest. Source S Preamble SD Length A address address P 802.3

Table 1-3

1 1-2 S S A Control P 802.2

802.3

3

2

OUI

Variable

4

TYPE Data FCS

SNAP

802.3

Ethernet Header Fields Field

Description

Preamble (DIX)

Provides synchronization and signal transitions to allow proper clocking of the transmitted signal. Consists of 62 alternating 1s and 0s, and ends with a pair of 1s.

Preamble and Start of Frame Delimiter (802.3)

Same purpose and binary value as DIX preamble; 802.3 simply renames the 8-byte DIX preamble as a 7-byte preamble and a 1-byte Start of Frame Delimiter (SFD).

Type (or Protocol Type) (DIX)

2-byte field that identifies the type of protocol or protocol header that follows the header. Allows the receiver of the frame to know how to process a received frame.

Length (802.3)

Describes the length, in bytes, of the data following the Length field, up to the Ethernet trailer. Allows an Ethernet receiver to predict the end of the received frame.

Destination Service Access Point (802.2)

DSAP; 1-byte protocol type field. The size limitations, along with other uses of the low-order bits, required the later addition of SNAP headers.

Source Service Access Point (802.2)

SSAP; 1-byte protocol type field that describes the upper-layer protocol that created the frame.

Ethernet Layer 2: Framing and Addressing

Table 1-3

15

Ethernet Header Fields (Continued) Field

Description

Control (802.2)

1- or 2-byte field that provides mechanisms for both connectionless and connection-oriented operation. Generally used only for connectionless operation by modern protocols, with a 1-byte value of 0x03.

Organizationally Unique Identifier (SNAP)

OUI; 3-byte field, generally unused today, providing a place for the sender of the frame to code the OUI representing the manufacturer of the Ethernet NIC.

Type (SNAP)

2-byte Type field, using same values as the DIX Type field, overcoming deficiencies with size and use of the DSAP field.

Types of Ethernet Addresses Ethernet addresses, also frequently called MAC addresses, are 6 bytes in length, typically listed in hexadecimal form. There are three main types of Ethernet address, as listed in Table 1-4. Table 1-4

Three Types of Ethernet/MAC Address Type of Ethernet/MAC Address

Description and Notes

Unicast

Fancy term for an address that represents a single LAN interface. The I/G bit, the most significant bit in the most significant byte, is set to 0.

Broadcast

An address that means “all devices that reside on this LAN right now.” Always a value of hex FFFFFFFFFFFF.

Multicast

A MAC address that implies some subset of all devices currently on the LAN. By definition, the I/G bit is set to 1.

Most engineers instinctively know how unicast and broadcast addresses are used in a typical network. When an Ethernet NIC needs to send a frame, it puts its own unicast address in the Source Address field of the header. If it wants to send the frame to a particular device on the LAN, the sender puts the other device’s MAC address in the Ethernet header’s Destination Address field. If the sender wants to send the frame to every device on the LAN, it sends the frame to the FFFF.FFFF.FFFF broadcast destination address. (A frame sent to the broadcast address is named a broadcast or broadcast frame, and frames sent to unicast MAC addresses are called unicasts or unicast frames.) Multicast Ethernet frames are used to communicate with a possibly dynamic subset of the devices on a LAN. The most common use for Ethernet multicast addresses involves the use of IP multicast. For example, if only 3 of 100 users on a LAN want to watch the same video stream using an IP multicast–based video application, the application can send a single multicast frame. The three interested devices prepare by listening for frames sent to a particular multicast Ethernet address,

16

Chapter 1: Ethernet Basics

processing frames destined for that address. Other devices may receive the frame, but they ignore its contents. Because the concept of Ethernet multicast is most often used today with IP multicast, most of the rest of the details of Ethernet multicast will be covered in Chapter 16, “Introduction to IP Multicasting.”

Ethernet Address Formats The IEEE intends for unicast addresses to be unique in the universe by administering the assignment of MAC addresses. The IEEE assigns each vendor a code to use as the first 3 bytes of its MAC addresses; that first half of the addresses is called the Organizationally Unique Identifier (OUI). The IEEE expects each manufacturer to use its OUI for the first 3 bytes of the MAC assigned to any Ethernet product created by that vendor. The vendor then assigns a unique value in the low-order 3 bytes for each Ethernet card that it manufactures—thereby ensuring global uniqueness of MAC addresses. Figure 1-5 shows the basic Ethernet address format, along with some additional details. Figure 1-5

Ethernet Address Format Least Significant Byte

Most Significant Byte

1st Byte

2nd Byte

3rd Byte

4th Byte

5th Byte

6th Byte

Vendor-Assigned

OUI

U/L I/G Bit Bit

1st Byte Least Significant Bit

Most Significant Bit

Note that Figure 1-5 shows the location of the most significant byte and most significant bit in each byte. IEEE documentation lists Ethernet addresses with the most significant byte on the left. However, inside each byte, the leftmost bit is the least significant bit, and the rightmost bit is the most significant bit. Many documents refer to the bit order as canonical; other documents refer to it as little-endian. Regardless of the term, the bit order inside each byte is important for understanding the meaning of the two most significant bits in an Ethernet address: ■

The Individual/Group (I/G) bit

■

The Universal/Local (U/L) bit

Table 1-5 summarizes the meaning of each bit.

Ethernet Layer 2: Framing and Addressing

Table 1-5

17

I/G and U/L Bits Field

Meaning

I/G

Binary 0 means the address is a unicast; Binary 1 means the address is a multicast or broadcast.

U/L

Binary 0 means the address is vendor assigned; Binary 1 means the address has been administratively assigned, overriding the vendor-assigned address.

The I/G bit signifies whether the address represents an individual device or a group of devices, and the U/L bit identifies locally configured addresses. For instance, the Ethernet multicast addresses used by IP multicast implementations always start with 0x01005E. Hex 01 (the first byte of the address) converts to binary 00000001, with the most significant bit being 1, confirming the use of the I/G bit. NOTE Often, when overriding the MAC address to use a local address, the device or device driver does not enforce the setting of the U/L bit to a value of 1.

Protocol Types and the 802.3 Length Field Each of the three types of Ethernet header shown in Figure 1-4 has a field identifying the format of the Data field in the frame. Generically called a Type field, these fields allow the receiver of an Ethernet frame to know how to interpret the data in the received frame. For instance, a router might want to know whether the frame contains an IP packet, an IPX packet, and so on. DIX and the revised IEEE framing use the Type field, also called the Protocol Type field. The originally-defined IEEE framing uses those same 2 bytes as a Length field. To distinguish the style of Ethernet header, the Ethernet Type field values begin at 1536, and the length of the Data field in an IEEE frame is limited to decimal 1500 or less. That way, an Ethernet NIC can easily determine whether the frame follows the DIX or original IEEE format. The original IEEE frame used a 1-byte Protocol Type field (DSAP) for the 802.2 LLC standard type field. It also reserved the high-order 2 bits for other uses, similar to the I/G and U/L bits in MAC addresses. As a result, there were not enough possible combinations in the DSAP field for the needs of the market—so the IEEE had to define yet another type field, this one inside an additional IEEE SNAP header. Table 1-6 summarizes the meaning of the three main Type field options with Ethernet. Table 1-6

Ethernet Type Fields Type Field

Description

Protocol Type

DIX V2 Type field; 2 bytes; registered values now administered by the IEEE

DSAP

802.2 LLC; 1 byte, with 2 high-order bits reserved for other purposes; registered values now administered by the IEEE

SNAP

SNAP header; 2 bytes; uses same values as Ethernet Protocol Type; signified by an 802.2 DSAP of 0xAA

18

Chapter 1: Ethernet Basics

Switching and Bridging Logic In this chapter so far, you have been reminded about the cabling details for Ethernet along with the formats and meanings of the fields inside Ethernet frames. A switch’s ultimate goal is to deliver those frames to the appropriate destination(s) based on the destination MAC address in the frame header. Table 1-7 summarizes the logic used by switches when forwarding frames, which differs based on the type of destination Ethernet address and on whether the destination address has been added to its MAC address table. LAN Switch Forwarding Behavior

Table 1-7

Type of Address

Switch Action

Known unicast

Forwards frame out the single interface associated with the destination address

Unknown unicast

Floods frame out all interfaces, except the interface on which the frame was received

Broadcast

Floods frame identically to unknown unicasts

Multicast

Floods frame identically to unknown unicasts, unless multicast optimizations are configured

For unicast forwarding to work most efficiently, switches need to know about all the unicast MAC addresses and out which interface the switch should forward frames sent to each MAC address. Switches learn MAC addresses, and the port to associate with them, by reading the source MAC address of received frames. You can see the learning process in Example 1-2, along with several other details of switch operation. Figure 1-6 lists the devices in the network associated with Example 1-2, along with their MAC addresses. Figure 1-6

Sample Network with MAC Addresses Shown VLAN 1: 0200.3333.3333

0/3

SW1

R3

IP Address 10.1.1.1 MAC Address 000a.b7dc.b78d

0/13 000a.b7dc.b78d 000f.2343.87cd 0/13 0200.4444.4444

R4

VLAN 1:

0/4

SW4

IP Address 10.1.1.4 MAC Address 000f.2343.87c0

0/6

0010.a49b.6111

PC1

Switching and Bridging Logic

Example 1-2

19

Command Output Showing MAC Address Table Learning

Switch1# show mac-address-table dynamic Mac Address Table -----------------------------------------Vlan

Mac Address

Type

Ports

----

-----------

----

-----

1

000f.2343.87cd

DYNAMIC

Fa0/13

1

0200.3333.3333

DYNAMIC

Fa0/3

1

0200.4444.4444

DYNAMIC

Fa0/13

Total Mac Addresses for this criterion: 3 ! Above, Switch1’s MAC address table lists three dynamically learned addresses, ! including Switch4’s FA 0/13 MAC. ! Below, Switch1 pings Switch4’s management IP address. Switch1# ping 10.1.1.4 Type escape sequence to abort. Sending 5, 100-byte ICMP Echos to 10.1.1.4, timeout is 2 seconds: !!!!! Success rate is 100 percent (5/5), round-trip min/avg/max = 1/2/4 ms ! Below Switch1 now knows the MAC address associated with Switch4’s management IP ! address. Each switch has a range of reserved MAC addresses, with the first MAC ! being used by the switch IP address, and the rest being assigned in sequence to ! the switch interfaces – note 0xcd (last byte of 2nd address in the table above) ! is for Switch4’s FA 0/13 interface, and is 13 (decimal) larger than Switch4’s ! base MAC address. Switch1# show mac-address-table dynamic Mac Address Table -----------------------------------------Vlan

Mac Address

Type

Ports

----

-----------

----

-----

1

000f.2343.87c0

DYNAMIC

Fa0/13

1

000f.2343.87cd

DYNAMIC

Fa0/13

1

0200.3333.3333

DYNAMIC

Fa0/3

1

0200.4444.4444

DYNAMIC

Fa0/13

Total Mac Addresses for this criterion: 4 ! Not shown: PC1 ping 10.1.1.23 (R3) PC1's MAC in its MAC address table -----------------------------------------Vlan

Mac Address

Type

Ports

----

-----------

----

-----

1

000f.2343.87c0

DYNAMIC

Fa0/13

1

000f.2343.87cd

DYNAMIC

Fa0/13

1

0010.a49b.6111

DYNAMIC

Fa0/13

continues

20

Chapter 1: Ethernet Basics

Example 1-2

Command Output Showing MAC Address Table Learning (Continued) 1

0200.3333.3333

DYNAMIC

Fa0/3

1

0200.4444.4444

DYNAMIC

Fa0/13

Total Mac Addresses for this criterion: 5 ! Above, Switch1 learned the PC’s MAC address, associated with FA 0/13, ! because the frames sent by the PC came into Switch1 over its FA 0/13. ! Below, Switch4’s MAC address table shows PC1’s MAC off its FA 0/6 switch4# show mac-address-table dynamic Mac Address Table ------------------------------------------Vlan

Mac Address

Type

Ports

----

-----------

--------

-----

1

000a.b7dc.b780

DYNAMIC

Fa0/13

1

000a.b7dc.b78d

DYNAMIC

Fa0/13

1

0010.a49b.6111

DYNAMIC

Fa0/6

1

0200.3333.3333

DYNAMIC

Fa0/13

1

0200.4444.4444

DYNAMIC

Fa0/4

Total Mac Addresses for this criterion: 5 ! Below, for example, the aging timeout (default 300 seconds) is shown, followed ! by a command just listing the mac address table entry for a single address. switch4# show mac-address-table aging-time Vlan

Aging Time

----

----------

1

300

switch4# show mac-address-table address 0200.3333.3333 Mac Address Table ------------------------------------------Vlan

Mac Address

Type

Ports

----

-----------

--------

-----

0200.3333.3333

DYNAMIC

Fa0/13

1

Total Mac Addresses for this criterion: 1

SPAN and RSPAN Cisco Catalyst switches support a method of directing all traffic from a source port or source VLAN to a single port. This feature, called SPAN (for Switch Port Analyzer) in the Cisco documentation and sometimes referred to as session monitoring because of the commands used to configure it, is useful for many applications. These include monitoring traffic for compliance reasons, data collection purposes, or to support a particular application. For example, all traffic from a voice VLAN can be delivered to a single switch port to facilitate call recording in a VoIP network. Another common use of this feature is to support intrusion detection/prevention system (IDS/IPS) security solutions.

SPAN and RSPAN

21

SPAN sessions can be sourced from a port or ports, or from a VLAN. This provides great flexibility in collecting or monitoring traffic from a particular source device or an entire VLAN. The destination port for a SPAN session can be on the local switch, as in SPAN operation. Or it can be a port on another switch in the network. This mode is known as Remote SPAN, or RSPAN. In RSPAN, a specific VLAN must be configured across the entire switching path from the source port or VLAN to the RSPAN destination port. This requires that the RSPAN VLAN be included in any trunks in that path, too. See Figure 1-7 for the topology of SPAN, and Figure 1-8 for that of RSPAN. Figure 1-7

SPAN Topology Catalyst Switch

SPAN Source VLAN or Port SPAN Destination Port

Figure 1-8

RSPAN Topology Switched Network

RSPAN Source VLAN(s) or Port(s)

RSPAN Destination Port RSPAN VLAN

The information in this section applies specifically to the Cisco 3560 switching platform; the Cisco 3750 and many other platforms use identical or similar rules and configuration commands.

22

Chapter 1: Ethernet Basics

Core Concepts of SPAN and RSPAN To understand SPAN and RSPAN, it helps to break them down into their fundamental elements. This also helps you understand configuring these features. In SPAN, you create a SPAN source that consists of at least one port or at least one VLAN on a switch. On the same switch, you configure a destination port. The SPAN source data is then gathered and delivered to the SPAN destination. In RSPAN, you create the same source type—at least one port or at least one VLAN. The destination for this session is the RSPAN VLAN, rather than a single port on the switch. At the switch that contains an RSPAN destination port, the RSPAN VLAN data is delivered to the RSPAN port. A SPAN source port can be any type of port—a routed port, a physical switch port, an access port, a trunk port, an EtherChannel port (either one physical port or the entire port-channel interface), and so on. On a SPAN source VLAN, all active ports in that VLAN are monitored. As you add or remove ports from that VLAN, the sources are dynamically updated to include new ports or exclude removed ports. Also, a port configured as a SPAN destination cannot be part of a SPAN source VLAN.

Restrictions and Conditions Destination ports in SPAN and RSPAN have multiple restrictions. The key restrictions include the following: ■

When you configure a destination port, its original configuration is overwritten. If the SPAN configuration is removed, the original configuration on that port is restored.

■

When you configure a destination port, the port is removed from any EtherChannel bundle if it were part of one. If it were a routed port, the SPAN destination configuration overrides the routed port configuration.

■

Destination ports do not support port security, 802.1x authentication, or private VLANs. In general, SPAN/RSPAN and 802.1x are incompatible.

■

Destination ports do not support any Layer 2 protocols, including CDP, Spanning Tree, VTP, DTP, and so on.

A set of similar restrictions for RSPAN destination VLANs also exists. See the references in the “Further Reading” section at the end of this chapter for more information about those restrictions.

SPAN and RSPAN

23

SPAN and RSPAN require compliance with a number of specific conditions to work. For SPAN, the key restrictions include the following: ■

The source can be either one or more ports or a VLAN, but not a mix of these.

■

Up to 64 SPAN destination ports can be configured on a switch.

■

Switched or routed ports can be configured as SPAN source ports or SPAN destination ports.

■

Be careful to avoid overloading the SPAN destination port. A 100-Mbps source port can easily overload a 10-Mbps destination port; it’s even easier to overload a 100-Mbps destination port when the source is a VLAN.

■

Within a single SPAN session, you cannot deliver traffic to a destination port when it is sourced by a mix of SPAN and RSPAN source ports or VLANs. This restriction comes into play when you want to mirror traffic to both a local port on a switch (in SPAN) and a remote port on another switch (in RSPAN mode).

■

A SPAN destination port cannot be a source port, and a source port cannot be a destination port.

■

Only one SPAN/RSPAN session can send traffic to a single destination port.

■

A SPAN destination port ceases to act as a normal switchport. That is, it passes only SPANrelated traffic.

■

It’s possible to configure a trunk port as the source of a SPAN or RSPAN session. In this case, all VLANs on the trunk are monitored by default; the filter vlan command option can be configured to limit the VLANs being monitored in this situation.

■

Traffic that is routed from another VLAN to a source VLAN cannot be monitored with SPAN. An easy way to understand this concept is that only traffic that enters or exits the switch in a source port or VLAN is forwarded in a SPAN session. In other words, if the traffic comes from another source within the switch (by routing from another VLAN, for example), that traffic isn’t forwarded via SPAN.

SPAN and RSPAN support two types of traffic: transmitted and received. By default, SPAN is enabled for traffic both entering and exiting the source port or VLAN. However, SPAN can be configured to monitor just transmitted traffic or just received traffic. Some additional conditions apply to these traffic types, as detailed in this list: ■

For Receive (RX) SPAN, the goal is to deliver all traffic received to the SPAN destination. As a result, each frame to be transported across a SPAN connection is copied and sent before any modification (for example, VACL or ACL filtering, QoS modification, or even ingress or egress policing).

24

Chapter 1: Ethernet Basics

■

For Transmit (TX) SPAN, all relevant filtering or modification by ACLs, VACLs, QoS, or policing actions are taken before the switch forwards the traffic to the SPAN/RSPAN destination. As a result, not all transmit traffic necessarily makes it to a SPAN destination. Also, the frames that are delivered do not necessarily match the original frames exactly, depending on policies applied before they are forwarded to the SPAN destination.

■

A special case applies to certain types of Layer 2 frames. SPAN/RSPAN usually ignores CDP, spanning-tree BPDUs, VTP, DTP, and PagP frames. However, these traffic types can be forwarded along with the normal SPAN traffic if the encapsulation replicate command is configured.

Basic SPAN Configuration The goal for the configuration in Example 1-3 is to mirror traffic sent or received from interface fa0/12 to interface fa0/24. All traffic sent or received on fa0/12 is sent to fa0/24. This configuration is typical of a basic traffic monitoring application. Example 1-3

Basic SPAN Configuration Example

MDF-ROC1# configure terminal MDF-ROC1(config)# monitor session 1 source interface fa0/12 MDF-ROC1(config)# monitor session 1 destination interface fa0/24

Complex SPAN Configuration In Example 1-4, we configure a switch to send the following traffic to interface fa0/24, preserving the encapsulation from the sources: ■

Received on interface fa0/18

■

Sent on interface fa0/9

■

Sent and received on interface fa0/19 (which is a trunk)

We also filter (remove) VLANs 1, 2, 3, and 229 from the traffic coming from the fa0/19 trunk port. Example 1-4

Complex SPAN Configuration Example

MDF-ROC3# config term MDF-ROC3(config)# monitor session 11 source interface fa0/18 rx MDF-ROC3(config)# monitor session 11 source interface fa0/9 tx MDF-ROC3(config)# monitor session 11 source interface fa0/19 MDF-ROC3(config)# monitor session 11 filter vlan 1 - 3 , 229 MDF-ROC3(config)# monitor session 11 destination interface fa0/24 encapsulation replicate

SPAN and RSPAN

25

RSPAN Configuration In Example 1-5, we configure two switches, IDF-SYR1 and IDF-SYR2, to send traffic to RSPAN VLAN 199, which is delivered to port fa0/24 on switch MDF-SYR9 as follows: ■

From IDF-SYR1, all traffic received on VLANs 66–68

■

From IDF-SYR2, all traffic received on VLAN 9

■

From IDF-SYR2, all traffic sent and received on VLAN 11

Note that all three switches use a different session ID, which is permissible in RSPAN. The only limitation on session numbering is that the session number must be 1 to 66.

Example 1-5

RSPAN Configuration Example

IDF-SYR1# config term IDF-SYR1(config)# vlan 199 IDF-SYR1(config-vlan)# remote span IDF-SYR1(config-vlan)# exit IDF-SYR1(config)# monitor session 3 source vlan 66 – 68 rx IDF-SYR1(config)# monitor session 3 destination remote vlan 199 !Now moving to IDF-SYR2: IDF-SYR2# config term IDF-SYR2(config)# vlan 199 IDF-SYR2(config-vlan)# remote span IDF-SYR2(config-vlan)# exit IDF-SYR2(config)# monitor session 23 source vlan 9 rx IDF-SYR2(config)# monitor session 23 source vlan 11 rx IDF-SYR2(config)# monitor session 23 destination remote vlan 199 !Now moving to MDF-SYR9 MDF-SYR9# config term MDF-SYR9(config)# vlan 199 MDF-SYR9(config-vlan)# remote span MDF-SYR9(config-vlan)# exit MDF-SYR9(config)# monitor session 63 source remote vlan 199 MDF-SYR9(config)# monitor session 63 destination interface fa0/24 MDF-SYR9(config)# end

You can verify SPAN or RSPAN operation using the show monitor session command. From a troubleshooting standpoint, it’s important to note that if the destination port is shut down, the SPAN instance won’t come up. Once you bring the port up, the SPAN session will follow.

26

Chapter 1: Ethernet Basics

Foundation Summary This section lists additional details and facts to round out the coverage of the topics in this chapter. Unlike most of the Cisco Press Exam Certification Guides, this “Foundation Summary” does not repeat information presented in the “Foundation Topics” section of the chapter. Please take the time to read and study the details in the “Foundation Topics” section of the chapter, as well as review items noted with a Key Topic icon. Table 1-8 lists the different types of Ethernet and some distinguishing characteristics of each type. Ethernet Standards

Table 1-8

Type of Ethernet

General Description

10BASE5

Commonly called “thick-net”; uses coaxial cabling

10BASE2

Commonly called “thin-net”; uses coaxial cabling

10BASE-T

First type of Ethernet to use twisted-pair cabling

DIX Ethernet Version 2

Layer 1 and Layer 2 specifications for original Ethernet, from Digital/Intel/ Xerox; typically called DIX V2

IEEE 802.3

Called MAC due to the name of the IEEE committee (Media Access Control); original Layer 1 and 2 specifications, standardized using DIX V2 as a basis

IEEE 802.2

Called LLC due to the name of the IEEE committee (Logical Link Control); Layer 2 specification for header common to multiple IEEE LAN specifications

IEEE 802.3u

IEEE standard for Fast Ethernet (100 Mbps) over copper and optical cabling; typically called FastE

IEEE 802.3z

Gigabit Ethernet over optical cabling; typically called GigE

IEEE 802.3ab

Gigabit Ethernet over copper cabling

Switches forward frames when necessary, and do not forward when there is no need to do so, thus reducing overhead. To accomplish this, switches perform three actions: ■

Learn MAC addresses by examining the source MAC address of each received frame

■

Decide when to forward a frame or when to filter (not forward) a frame, based on the destination MAC address

■

Create a loop-free environment with other bridges by using the Spanning Tree Protocol

The internal processing algorithms used by switches vary among models and vendors; regardless, the internal processing can be categorized as one of the methods listed in Table 1-9.

Foundation Summary

27

Switch Internal Processing

Table 1-9

Switching Method

Description

Store-and-forward

The switch fully receives all bits in the frame (store) before forwarding the frame (forward). This allows the switch to check the FCS before forwarding the frame, thus ensuring that errored frames are not forwarded.

Cut-through

The switch performs the address table lookup as soon as the Destination Address field in the header is received. The first bits in the frame can be sent out the outbound port before the final bits in the incoming frame are received. This does not allow the switch to discard frames that fail the FCS check, but the forwarding action is faster, resulting in lower latency.

Fragment-free

This performs like cut-through switching, but the switch waits for 64 bytes to be received before forwarding the first bytes of the outgoing frame. According to Ethernet specifications, collisions should be detected during the first 64 bytes of the frame, so frames that are in error because of a collision will not be forwarded.

Table 1-10 lists some of the most popular Cisco IOS commands related to the topics in this chapter. Table 1-10

Catalyst IOS Commands for Catalyst Switch Configuration Command

Description

interface vlan 1

Global command; moves user to interface configuration mode for a VLAN interface

interface fastethernet 0/x

Puts user in interface configuration mode for that interface

duplex {auto | full | half}

Used in interface configuration mode; sets duplex mode for the interface

speed {10 | 100 | 1000 | auto | nonegotiate}

Used in interface configuration mode; sets speed for the interface

show mac address-table [agingtime | count | dynamic | static] [address hw-addr] [interface interface-id] [vlan vlan-id]

Displays the MAC address table; the security option displays information about the restricted or static settings

show interface fastethernet 0/x

Displays interface status for a physical 10/100 interface

show interface vlan 1

Displays IP address configuration for VLAN

remote span

In VLAN configuration mode, specifies that the VLAN is configured as a remote SPAN destination VLAN

monitor session 1-66 source [vlan vlan-id | interface interface-id] [rx | tx | both]

Configures a SPAN or RSPAN source, which can include one or more physical interfaces or one or more VLANs; optionally specifies traffic entering (Rx) or leaving (Tx), or both, with respect to the specified source continues

28

Chapter 1: Ethernet Basics

Table 1-10

Catalyst IOS Commands for Catalyst Switch Configuration (Continued) Command

Description

monitor session 1-66 destination [remote vlan vlan-id] | interface interface-id]

Configures the destination of a SPAN or RSPAN session to be either a physical interface or a remote VLAN

monitor session 1-66 filter vlan [vlan | vlan-range]

Removes traffic from the specified VLAN or VLAN range from the monitored traffic stream

show monitor session session-id

Displays the status of a SPAN session

Table 1-11 outlines the types of UTP cabling. Table 1-11

UTP Cabling Reference UTP Category

Max Speed Rating

Description

1

—

Used for telephones, and not for data

2

4 Mbps

Originally intended to support Token Ring over UTP

3

10 Mbps

Can be used for telephones as well; popular option for Ethernet in years past, if Cat 3 cabling for phones was already in place

4

16 Mbps

Intended for the fast Token Ring speed option

5

1 Gbps

Very popular for cabling to the desktop

5e

1 Gbps

Added mainly for the support of copper cabling for Gigabit Ethernet

6

1 Gbps+

Intended as a replacement for Cat 5e, with capabilities to support multigigabit speeds

Table 1-12 lists the pertinent details of the Ethernet standards and the related cabling. Table 1-12

Ethernet Types and Cabling Standards Standard

Cabling

Maximum Single Cable Length

10BASE5

Thick coaxial

500 m

10BASE2

Thin coaxial

185 m

10BASE-T

UTP Cat 3, 4, 5, 5e, 6

100 m

100BASE-FX

Two strands, multimode

400 m

100BASE-T

UTP Cat 3, 4, 5, 5e, 6, 2 pair

100 m

100BASE-T4

UTP Cat 3, 4, 5, 5e, 6, 4 pair

100 m

100BASE-TX

UTP Cat 3, 4, 5, 5e, 6, or STP, 2 pair

100 m

Memory Builders

Table 1-12

29

Ethernet Types and Cabling Standards (Continued) Standard

Cabling

Maximum Single Cable Length

1000BASE-LX

Long-wavelength laser, MM or SM fiber

10 km (SM) 3 km (MM)

1000BASE-SX

Short-wavelength laser, MM fiber

220 m with 62.5-micron fiber; 550 m with 50-micron fiber

1000BASE-ZX

Extended wavelength, SM fiber

100 km

1000BASE-CS

STP, 2 pair

25 m

1000BASE-T

UTP Cat 5, 5e, 6, 4 pair

100 m

Memory Builders The CCIE Routing and Switching written exam, like all Cisco CCIE written exams, covers a fairly broad set of topics. This section provides some basic tools to help you exercise your memory about some of the broader topics covered in this chapter.

Fill In Key Tables from Memory Appendix G, “Key Tables for CCIE Study,” on the CD in the back of this book contains empty sets of some of the key summary tables in each chapter. Print Appendix G, refer to this chapter’s tables in it, and fill in the tables from memory. Refer to Appendix H, “Solutions for Key Tables for CCIE Study,” on the CD to check your answers.

Definitions Next, take a few moments to write down the definitions for the following terms: Auto-negotiation, half duplex, full duplex, cross-over cable, straight-through cable, unicast address, multicast address, broadcast address, loopback circuitry, I/G bit, U/L bit, CSMA/CD, SPAN, RSPAN, remote VLAN, monitor session, VLAN filtering, encapsulation replication Refer to the glossary to check your answers.

Further Reading For a good reference for more information on the actual FLPs used by auto-negotiation, refer to the Fast Ethernet web page of the University of New Hampshire Research Computing Center’s InterOperability Laboratory, at http://www.iol.unh.edu/services/testing/fe/training/. For information about configuring SPAN and RSPAN, and for a full set of restrictions (specific to the 3560 and 3750), see http://www.ciscosystems.com/en/US/docs/switches/lan/catalyst3560/ software/release/12.2_50_se/configuration/guide/swspan.html.

Blueprint topics covered in this chapter: This chapter covers the following subtopics from the Cisco CCIE Routing and Switching written exam blueprint. Refer to the full blueprint in Table I-1 in the Introduction for more details on the topics covered in each chapter and their context within the blueprint. ■

VLANs

■

VLAN Trunking Protocol (VTP)

■

Etherchannel

■

PPP over Ethernet (PPPoE)

CHAPTER

2

Virtual LANs and VLAN Trunking This chapter continues with the coverage of some of the most fundamental and important LAN topics with coverage of VLANs and VLAN trunking. As usual, for those of you current in your knowledge of the topics in this chapter, review the items next to the Key Topic icons spread throughout the chapter, plus the “Foundation Summary” and “Memory Builders” sections at the end of the chapter.

“Do I Know This Already?” Quiz Table 2-1 outlines the major headings in this chapter and the corresponding “Do I Know This Already?” quiz questions. Table 2-1

“Do I Know This Already?” Foundation Topics Section-to-Question Mapping Foundation Topics Section

Questions Covered in This Section

Virtual LANs

1–2

VLAN Trunking Protocol

3–5

VLAN Trunking: ISL and 802.1Q

6–9

Configuring PPPoE

10

Score

Total Score

In order to best use this pre-chapter assessment, remember to score yourself strictly. You can find the answers in Appendix A, “Answers to the ‘Do I Know This Already?’ Quizzes.” 1.

Assume that VLAN 28 does not yet exist on Switch1. Which of the following commands, issued from any part of global configuration mode (reached with the configure terminal exec command) would cause the VLAN to be created? a.

vlan 28

b.

vlan 28 name fred

c.

switchport vlan 28

d.

switchport access vlan 28

e.

switchport access 28

32

Chapter 2: Virtual LANs and VLAN Trunking

2.

3.

4.

Which of the following are the two primary motivations for using private VLANs? a.

Better LAN security

b.

IP subnet conservation

c.

Better consistency in VLAN configuration details

d.

Reducing the impact of broadcasts on end-user devices

e.

Reducing the unnecessary flow of frames to switches that do not have any ports in the VLAN to which the frame belongs

Which of the following VLANs can be pruned by VTP on an 802.1Q trunk? a.

1–1023

b.

1–1001

c.

2–1001

d.

1–1005

e.

2–1005

An existing switched network has ten switches, with Switch1 and Switch2 being the only VTP servers in the network. The other switches are all VTP clients and have successfully learned about the VLANs from the VTP servers. The only configured VTP parameter on all switches is the VTP domain name (Larry). The VTP revision number is 201. What happens when a new, already-running VTP client switch, named Switch11, with domain name Larry and revision number 301, connects via a trunk to any of the other ten switches? a.

No VLAN information changes; Switch11 ignores the VTP updates sent from the two existing VTP servers until the revision number reaches 302.

b.

The original ten switches replace their old VLAN configuration with the configuration in Switch11.

c.

Switch11 replaces its own VLAN configuration with the configuration sent to it by one of the original VTP servers.

d.

Switch11 merges its existing VLAN database with the database learned from the VTP servers, because Switch11 had a higher revision number.

“Do I Know This Already?” Quiz

5.

6.

7.

33

An existing switched network has ten switches, with Switch1 and Switch2 being the only VTP servers in the network. The other switches are all VTP clients, and have successfully learned about the VLANs from the VTP server. The only configured VTP parameter is the VTP domain name (Larry). The VTP revision number is 201. What happens when an alreadyrunning VTP server switch, named Switch11, with domain name Larry and revision number 301, connects via a trunk to any of the other ten switches? a.

No VLAN information changes; all VTP updates between the original VTP domain and the new switch are ignored.

b.

The original ten switches replace their old VLAN configuration with the configuration in Switch11.

c.

Switch11 replaces its old VLAN configuration with the configuration sent to it by one of the original VTP servers.

d.

Switch11 merges its existing VLAN database with the database learned from the VTP servers, because Switch11 had a higher revision number.

e.

None of the other answers is correct.

Assume that two brand-new Cisco switches were removed from their cardboard boxes. PC1 was attached to one switch, PC2 was attached to the other, and the two switches were connected with a cross-over cable. The switch connection dynamically formed an 802.1Q trunk. When PC1 sends a frame to PC2, how many additional bytes of header are added to the frame before it passes over the trunk? a.

0

b.

4

c.

8

d.

26

Assume that two brand-new Cisco Catalyst 3550 switches were connected with a cross-over cable. Before attaching the cable, one switch interface was configured with the switchport trunk encapsulation dot1q, switchport mode trunk, and switchport nonegotiate subcommands. Which of the following must be configured on the other switch before trunking will work between the switches? a.

switchport trunk encapsulation dot1q

b.

switchport mode trunk

c.

switchport nonegotiate

d.

No configuration is required.

34

Chapter 2: Virtual LANs and VLAN Trunking

8.

9.

10.

When configuring trunking on a Cisco router fa0/1 interface, under which configuration modes could the IP address associated with the native VLAN (VLAN 1 in this case) be configured? a.

Interface fa 0/1 configuration mode

b.

Interface fa 0/1.1 configuration mode

c.

Interface fa 0/1.2 configuration mode

d.

None of the other answers is correct

Which of the following is false about 802.1Q? a.

Encapsulates the entire frame inside an 802.1Q header and trailer

b.

Supports the use of a native VLAN

c.

Allows VTP to operate only on extended-range VLANs

d.

Is chosen over ISL by DTP

Which command enables PPPoE on the outside Ethernet interface on a Cisco router? a.

pppoe enable

b.

pppoe-client enable

c.

pppoe-client dialer-pool-number

d.

pppoe-client dialer-number

Virtual LANs

35

Foundation Topics

Virtual LANs In an Ethernet LAN, a set of devices that receive a broadcast sent by any one of the devices in the same set is called a broadcast domain. On switches that have no concept of virtual LANs (VLAN), a switch simply forwards all broadcasts out all interfaces, except the interface on which it received the frame. As a result, all the interfaces on an individual switch are in the same broadcast domain. Also, if the switch connects to other switches and hubs, the interfaces on those switches and hubs are also in the same broadcast domain. A VLAN is simply an administratively defined subset of switch ports that are in the same broadcast domain. Ports can be grouped into different VLANs on a single switch, and on multiple interconnected switches as well. By creating multiple VLANs, the switches create multiple broadcast domains. By doing so, a broadcast sent by a device in one VLAN is forwarded to the other devices in that same VLAN; however, the broadcast is not forwarded to devices in the other VLANs. With VLANs and IP, best practices dictate a one-to-one relationship between VLANs and IP subnets. Simply put, the devices in a single VLAN are typically also in the same single IP subnet. Alternately, it is possible to put multiple subnets in one VLAN, and use secondary IP addresses on routers to route between the VLANs and subnets. Also, although not typically done, you can design a network to use one subnet on multiple VLANs, and use routers with proxy ARP enabled to forward traffic between hosts in those VLANs. (Private VLANs might be considered to consist of one subnet over multiple VLANs as well, as covered later in this chapter.) Ultimately, the CCIE written exams tend to focus more on the best use of technologies, so this book will assume that one subnet sits on one VLAN, unless otherwise stated. Layer 2 switches forward frames between devices in the same VLAN, but they do not forward frames between two devices in different VLANs. To forward data between two VLANs, a multilayer switch (MLS) or router is needed. Chapter 6, “IP Forwarding (Routing),” covers the details of MLS.

VLAN Configuration Configuring VLANs in a network of Cisco switches requires just a few simple steps: Step 1

Create the VLAN itself.

Step 2

Associate the correct ports with that VLAN.

36

Chapter 2: Virtual LANs and VLAN Trunking

The challenge relates to how some background tasks differ depending on how the Cisco VLAN Trunking Protocol (VTP) is configured, and whether normal-range or extended-range VLANs are being used. Using VLAN Database Mode to Create VLANs To begin, consider Example 2-1, which shows some of the basic mechanics of VLAN creation in VLAN database configuration mode. VLAN database configuration mode allows the creation of VLANs, basic administrative settings for each VLAN, and verification of VTP configuration information. Only normal-range (VLANs 1–1005) VLANs can be configured in this mode, and the VLAN configuration is stored in a Flash file called vlan.dat. Example 2-1 demonstrates VLAN database configuration mode, showing the configuration on Switch3 from Figure 2-1. The example shows VLANs 21 and 22 being created. Simple Access and Distribution

Figure 2-1

VLAN 21

VLAN 22

Subnet 10.1.21.x/24

Subnet 10.1.22.x/24

Fa0/0 Fa0/1

Gi0/1

SW1

R1

Fa0/12

Fa0/2

SW2

R2

Fa0/24

Fa0/3

R3

SW3

SW4

Fa0/7

Fa0/5

PC1

Example 2-1

R4

VLAN Creation in VLAN Database Mode–Switch3

! Below, note that FA 0/12 and FA0/24 missing from the list, because they have ! dynamically become trunks, supporting multiple VLANs. Switch3# show vlan brief VLAN Name

Status

Ports

---- -------------------------------- --------- ------------------------------1

default

active

Fa0/1, Fa0/2, Fa0/3, Fa0/4 Fa0/5, Fa0/6, Fa0/7, Fa0/8 Fa0/9, Fa0/10, Fa0/11, Fa0/13 Fa0/14, Fa0/15, Fa0/16, Fa0/17 Fa0/18, Fa0/19, Fa0/20, Fa0/21 Fa0/22, Fa0/23

Virtual LANs

Example 2-1

37

VLAN Creation in VLAN Database Mode–Switch3 (Continued)

! Below, “unsup” means that this 2950 switch does not support FDDI and TR 1002 fddi-default

act/unsup

1003 token-ring-default

act/unsup

1004 fddinet-default

act/unsup

1005 trnet-default

act/unsup

! Below, vlan database moves user to VLAN database configuration mode. ! The vlan 21 command defines the VLAN, as seen in the next command output show current), VLAN 21 is not in the “current” VLAN list. ! (s Switch3# vlan database Switch3(vlan)# vlan 21 VLAN 21 added: Name: VLAN0021 ! The show current command lists the VLANs available to the IOS when the switch ! is in VTP Server mode. The command lists the VLANs in numeric order, with ! VLAN 21 missing. Switch3(vlan)# show current VLAN ISL Id: 1 Name: default Media Type: Ethernet VLAN 802.10 Id: 100001 State: Operational MTU: 1500 Backup CRF Mode: Disabled Remote SPAN VLAN: No VLAN ISL Id: 1002 Name: fddi-default Media Type: FDDI VLAN 802.10 Id: 101002 State: Operational MTU: 1500 Backup CRF Mode: Disabled Remote SPAN VLAN: No ! Lines omitted for brevity ! Next, note that show proposed lists VLAN 21. The vlan 21 command ! creates the definition, but it must be “applied” before it is “current”. Switch3(vlan)# show proposed VLAN ISL Id: 1 Name: default Media Type: Ethernet VLAN 802.10 Id: 100001 State: Operational MTU: 1500 Backup CRF Mode: Disabled Remote SPAN VLAN: No

continues

38

Chapter 2: Virtual LANs and VLAN Trunking

Example 2-1

VLAN Creation in VLAN Database Mode–Switch3 (Continued) VLAN ISL Id: 21 Name: VLAN0021 Media Type: Ethernet VLAN 802.10 Id: 100021 State: Operational MTU: 1500 Backup CRF Mode: Disabled Remote SPAN VLAN: No

! Lines omitted for brevity ! Next, you could apply to complete the addition of VLAN 21, ! abort to not make the changes and exit VLAN database mode, or ! reset to not make the changes but stay in VLAN database mode. Switch3(vlan)# ? VLAN database editing buffer manipulation commands: abort

Exit mode without applying the changes

apply

Apply current changes and bump revision number

exit

Apply changes, bump revision number, and exit mode

no

Negate a command or set its defaults

reset

Abandon current changes and reread current database

show

Show database information

vlan

Add, delete, or modify values associated with a single VLAN

vtp

Perform VTP administrative functions.

! The apply command was used, making the addition of VLAN 21 complete. Switch3(vlan)# apply APPLY completed. ! A show current now would list VLAN 21. Switch3(vlan)# vlan 22 name ccie-vlan-22 VLAN 22 added: Name: ccie-vlan-22 ! Above and below, some variations on commands are shown, along with the ! creation of VLAN 22, with name ccie-vlan-22. ! Below, the vlan 22 option is used on show current and show proposed ! detailing the fact that the apply has not been done yet. Switch3(vlan)# show current 22 VLAN 22 does not exist in current database Switch3(vlan)# show proposed 22 VLAN ISL Id: 22 ! Lines omitted for brevity ! Finally, the user exits VLAN database mode using CTRL-Z, which does ! not inherently apply the change. CTRL-Z actually executes an abort. Switch3(vlan)# ^Z

Using Configuration Mode to Put Interfaces into VLANs To make a VLAN operational, the VLAN must be created, and then switch ports must be assigned to the VLAN. Example 2-2 shows how to associate the interfaces with the correct VLANs, once again on Switch3.

Virtual LANs

39

NOTE At the end of Example 2-1, VLAN 22 had not been successfully created. The assumption for Example 2-2 is that VLAN 22 has been successfully created.

Example 2-2

Assigning Interfaces to VLANs–Switch3

! First, the switchport access command assigns the VLAN numbers to the ! respective interfaces. Switch3# config t Enter configuration commands, one per line.

End with CNTL/Z.

Switch3(config)# int fa 0/3 Switch3(config-if)# switchport access vlan 22 Switch3(config-if)# int fa 0/7 Switch3(config-if)# switchport access vlan 21 Switch3(config-if)# ^Z ! Below, show vlan brief lists these same two interfaces as now being in ! VLANs 21 and 22, respectively. Switch3# show vlan brief VLAN Name

Status

Ports

---- -------------------------------- --------- ------------------------------1

default

active

Fa0/1, Fa0/2, Fa0/4, Fa0/5 Fa0/6, Fa0/8, Fa0/9, Fa0/10 Fa0/11, Fa0/13, Fa0/14, Fa0/15 Fa0/16, Fa0/17, Fa0/18, Fa0/19 Fa0/20, Fa0/21, Fa0/22, Fa0/23

21

VLAN0021

active

Fa0/7

22

ccie-vlan-22

active

Fa0/3

! Lines omitted for brevity ! While the VLAN configuration is not shown in the running-config at this point, ! the switchport access command that assigns the VLAN for the interface is in the ! configuration, as seen with the show run int fa 0/3 command. Switch3# show run int fa 0/3 interface FastEthernet0/3 switchport access vlan 22

Using Configuration Mode to Create VLANs At this point, the two new VLANs (21 and 22) have been created on Switch3, and the two interfaces are now in the correct VLANs. However, Cisco IOS switches support a different way to create VLANs, using configuration mode, as shown in Example 2-3. Example 2-3

Creating VLANs in Configuration Mode–Switch3

! First, VLAN 31 did not exist when the switchport access vlan 31 command was ! issued. As a result, the switch both created the VLAN and put interface fa0/8 ! into that VLAN. Then, the vlan 32 global command was used to create a

continues

40

Chapter 2: Virtual LANs and VLAN Trunking

Example 2-3

Creating VLANs in Configuration Mode–Switch3 (Continued)

! VLAN from configuration mode, and the name subcommand was used to assign a ! non-default name. Switch3# conf t Enter configuration commands, one per line.

End with CNTL/Z.

Switch3(config)# int fa 0/8 Switch3(config-if)# switchport access vlan 31 % Access VLAN does not exist. Creating vlan 31 Switch3(config-if)# exit Switch3(config)# vlan 32 Switch3(config-vlan)# name ccie-vlan-32 Switch3(config-vlan)# ^Z Switch3# show vlan brief VLAN Name

Status

Ports

---- -------------------------------- --------- ------------------------------1

default

active

Fa0/1, Fa0/2, Fa0/4, Fa0/5 Fa0/6, Fa0/9, Fa0/10, Fa0/11 Fa0/13, Fa0/14, Fa0/15, Fa0/16 Fa0/17, Fa0/18, Fa0/19, Fa0/20 Fa0/21, Fa0/22, Fa0/23

21

VLAN0021

active

Fa0/7

22

ccie-vlan-22

active

Fa0/3

31

VLAN0031

active

Fa0/8

32

ccie-vlan-32

active

! Portions omitted for brevity

Example 2-3 shows how the switchport access vlan subcommand creates the VLAN, as needed, and assigns the interface to that VLAN. Note that in Example 2-3, the show vlan brief output lists fa0/8 as being in VLAN 31. Because no ports have been assigned to VLAN 32 as of yet, the final line in Example 2-3 simply does not list any interfaces. The VLAN creation process is simple but laborious in a large network. If many VLANs exist, and they exist on multiple switches, instead of manually configuring the VLANs on each switch, you can use VTP to distribute the VLAN configuration of a VLAN to the rest of the switches. VTP will be discussed after a brief discussion of private VLANs.

Private VLANs Engineers may design VLANs with many goals in mind. In many cases today, devices end up in the same VLAN just based on the physical locations of the wiring drops. Security is another motivating factor in VLAN design: devices in different VLANs do not overhear each other’s

Virtual LANs

41

broadcasts. Additionally, the separation of hosts into different VLANs and subnets requires an intervening router or multilayer switch between the subnets, and these types of devices typically provide more robust security features. Regardless of the design motivations behind grouping devices into VLANs, good design practices typically call for the use of a single IP subnet per VLAN. In some cases, however, the need to increase security by separating devices into many small VLANs conflicts with the design goal of conserving the use of the available IP subnets. The Cisco private VLAN feature addresses this issue. Private VLANs allow a switch to separate ports as if they were on different VLANs, while consuming only a single subnet. A common place to implement private VLANs is in the multitenant offerings of a service provider (SP). The SP can install a single router and a single switch. Then, the SP attaches devices from multiple customers to the switch. Private VLANs then allow the SP to use only a single subnet for the whole building, separating different customers’ switch ports so that they cannot communicate directly, while supporting all customers with a single router and switch. Conceptually, a private VLAN includes the following general characterizations of how ports communicate: ■

Ports that need to communicate with all devices

■

Ports that need to communicate with each other, and with shared devices, typically routers

■

Ports that need to communicate only with shared devices

To support each category of allowed communications, a single private VLAN features a primary VLAN and one or more secondary VLANs. The ports in the primary VLAN are promiscuous in that they can send and receive frames with any other port, including ports assigned to secondary VLANs. Commonly accessed devices, such as routers and servers, are placed into the primary VLAN. Other ports, such as customer ports in the SP multitenant model, attach to one of the secondary VLANs. Secondary VLANs are either community VLANs or isolated VLANs. The engineer picks the type based on whether the device is part of a set of ports that should be allowed to send frames back and forth (community VLAN ports), or whether the device port should not be allowed to talk to any other ports besides those on the primary VLAN (isolated VLAN). Table 2-2 summarizes the behavior of private VLAN communications between ports.

42

Chapter 2: Virtual LANs and VLAN Trunking

Table 2-2

Private VLAN Communications Between Ports Description of Who Can Talk to Whom

Primary VLAN Ports

Community VLAN Ports1

Isolated VLAN Ports1

Talk to ports in primary VLAN (promiscuous ports)

Yes

Yes

Yes

Talk to ports in the same secondary VLAN (host ports)

N/A2

Yes

No

Talks to ports in another secondary VLAN

N/A2

No

No

1Community

and isolated VLANs are secondary VLANs.

2Promiscuous

ports, by definition in the primary VLAN, can talk to all other ports.

VLAN Trunking Protocol VTP advertises VLAN configuration information to neighboring switches so that the VLAN configuration can be made on one switch, with all the other switches in the network learning the VLAN information dynamically. VTP advertises the VLAN ID, VLAN name, and VLAN type for each VLAN. However, VTP does not advertise any information about which ports (interfaces) should be in each VLAN, so the configuration to associate a switch interface with a particular VLAN (using the switchport access vlan command) must still be configured on each individual switch. Also, the existence of the VLAN IDs used for private VLANs is advertised, but the rest of the detailed private VLAN configuration is not advertised by VTP. Each Cisco switch uses one of three VTP modes, as outlined in Table 2-3. Table 2-3

VTP Modes and Features* Function

Server Mode

Client Mode

Transparent Mode

Originates VTP advertisements

Yes

Yes

No

Processes received advertisements to update its VLAN configuration

Yes

Yes

No

Forwards received VTP advertisements

Yes

Yes

Yes

Saves VLAN configuration in NVRAM or vlan.dat

Yes

Yes

Yes

Can create, modify, or delete VLANs using configuration commands

Yes

No

Yes

*CatOS switches support a fourth VTP mode (off), meaning that the switch does not create, listen to, or forward VTP updates.

VLAN Trunking Protocol

43

VTP Process and Revision Numbers The VTP update process begins when a switch administrator, from a VTP server switch, adds, deletes, or updates the configuration for a VLAN. When the new configuration occurs, the VTP server increments the old VTP revision number by 1, and advertises the entire VLAN configuration database along with the new revision number. The VTP revision number concept allows switches to know when VLAN database changes have occurred. Upon receiving a VTP update, if the revision number in a received VTP update is larger than a switch’s current revision number, it believes that there is a new version of the VLAN database. Figure 2-2 shows an example in which the old VTP revision number was 3, the server adds a new VLAN, incrementing the revision number to 4, and then propagates the VTP database to the other switches. Figure 2-2

VTP Revision Number Basic Operation 1 Add New VLAN 2 Rev 3

3 Send VTP Advertisement VTP client

Rev 4 4 Rev 3 5 Sync New VLAN Info

VTP Server

Rev 4

3 Send VTP Advertisement VTP Client 4 Rev 3 Rev 4 5 Sync New VLAN Info

Cisco switches default to use VTP server mode, but they do not start sending VTP updates until the switch has been configured with a VTP domain name. At that point, the server begins to send its VTP updates, with a different database and revision number each time its VLAN configuration changes. However, the VTP clients in Figure 2-2 actually do not have to have the VTP domain name configured. If not configured, the client will assume it should use the VTP domain name in the first received VTP update. However, the client does need one small bit of configuration, namely, the VTP mode, as configured with the vtp mode global configuration command. VTP clients and servers alike will accept VTP updates from other VTP server switches. When using VTP, for better availability, a switched network using VTP needs at least two VTP server switches. Under normal operations, a VLAN change could be made on one server switch, and the other VTP server (plus all the clients) would learn about the changes to the VLAN database. Once learned, both VTP servers and clients store the VLAN configuration in their respective vlan.dat files in flash memory; they do not store the VLAN configuration in NVRAM.

44

Chapter 2: Virtual LANs and VLAN Trunking

With multiple VTP servers installed in a LAN, it is possible to accidentally overwrite the VTP configuration in the network. If trunks fail and then changes are made on more than one VTP server, the VTP configuration databases could differ, with different configuration revision numbers. When the formerly-separated parts of the LAN reconnect using trunks, the VTP database with a higher revision number is propagated throughout the VTP domain, replacing some switches’ VTP databases. Note also that because VTP clients can actually originate VTP updates, under the right circumstances, a VTP client can update the VTP database on another VTP client or server. See http://www.ciscopress.com/ 1587201968 and look for downloads, to download a document that describes how a client could update the VLAN database on another VTP client or server. In summary, for a newly-connected VTP server or client to change another switch’s VTP database, the following must be true: ■

The new link connecting the new switch is trunking.

■

The new switch has the same VTP domain name as the other switches.

■

The new switch’s revision number is larger than that of the existing switches.

■

The new switch must have the same password, if configured on the existing switches.

The revision number and VTP domain name can be easily seen with a Sniffer trace; to prevent DoS attacks with VTP, set VTP passwords, which are encoded as message digests (MD5) in the VTP updates. Also, some installations simply use VTP transparent mode on all switches, which prevents switches from ever listening to other switch VTP updates and erroneously deleting their VLAN configuration databases.

VTP Configuration VTP sends updates out all active trunk interfaces (ISL or 802.1Q). However, with all default settings from Cisco, switches are in server mode, with no VTP domain name configured, and they do not send any VTP updates. Before any switches can learn VLAN information from another switch, at least one switch must have a bare-minimum VTP server configuration—specifically, a domain name. Example 2-4 shows Switch3 configuring a VTP domain name to become a VTP server and advertise the VLANs it has configured. The example also lists several key VTP show commands. (Note that the example begins with VLANs 21 and 22 configured on Switch3, and all default settings for VTP on all four switches.) Example 2-4

VTP Configuration and show Command Example

! First, Switch3 is configured with a VTP domain ID of CCIE-domain. Switch3# conf t Enter configuration commands, one per line.

End with CNTL/Z.

Switch3(config)# vtp domain CCIE-domain Changing VTP domain name from NULL to CCIE-domain Switch3(config)# ^Z ! Next, on Switch1, the VTP status shows the same revision as Switch3, and it ! learned the VTP domain name CCIE-domain. Note that Switch1 has no VTP-related

VLAN Trunking Protocol

Example 2-4

45

VTP Configuration and show Command Example (Continued)

! configuration, so it is a VTP server; it learned the VTP domain name from. ! Switch3. Switch1# sh vtp status VTP Version

: 2

Configuration Revision

: 2

Maximum VLANs supported locally : 1005 Number of existing VLANs

: 7

VTP Operating Mode

: Server

VTP Domain Name

: CCIE-domain

VTP Pruning Mode

: Disabled

VTP V2 Mode

: Disabled

VTP Traps Generation

: Disabled

MD5 digest

: 0x0E 0x07 0x9D 0x9A 0x27 0x10 0x6C 0x0B

Configuration last modified by 10.1.1.3 at 3-1-93 00:02:55 Local updater ID is 10.1.1.1 on interface Vl1 (lowest numbered VLAN interface found) ! The show vlan brief command lists the VLANs learned from Switch3. Switch1# show vlan brief VLAN Name

Status

Ports

---- -------------------------------- --------- ------------------------------1

default

active

Fa0/1, Fa0/2, Fa0/3, Fa0/4 Fa0/5, Fa0/6, Fa0/7, Fa0/10 Fa0/11, Fa0/13, Fa0/14, Fa0/15 Fa0/16, Fa0/17, Fa0/18, Fa0/19 Fa0/20, Fa0/21, Fa0/22, Fa0/23 Gi0/2

21

VLAN0021

active

22

ccie-vlan-22

active

1002 fddi-default

active

1003 token-ring-default

active

1004 fddinet-default

active

1005 trnet-default

active

Example 2-4 shows examples of a few VTP configuration options. Table 2-4 provides a complete list, along with explanations. Table 2-4

VTP Configuration Options Option

Meaning

domain

Sends domain name in VTP updates. Received VTP update is ignored if it does not match a switch’s domain name. One VTP domain name per switch is allowed.

password

Used to generate an MD5 hash that is included in VTP updates. Received VTP updates are ignored if the passwords on the sending and receiving switch do not match.

mode

Sets server, client, or transparent mode on the switch. continues

46

Chapter 2: Virtual LANs and VLAN Trunking

Table 2-4

VTP Configuration Options (Continued) Option

Meaning

version

Sets version 1 or 2. Servers and clients must match version to exchange VLAN configuration data. Transparent mode switches at version 2 forward version 1 or version 2 VTP updates.

pruning

Enables VTP pruning, which prevents flooding on a per-VLAN basis to switches that do not have any ports configured as members of that VLAN.

interface

Specifies the interface whose IP address is used to identify this switch in VTP updates.

Normal-Range and Extended-Range VLANs Some VLAN numbers are considered to be normal, whereas some others are considered to be extended. Normal-range VLANs are VLANs 1–1005, and can be advertised via VTP versions 1 and 2. These VLANs can be configured in VLAN database mode, with the details being stored in the vlan.dat file in Flash. Extended-range VLANs range from 1006–4094, inclusive. However, these additional VLANs cannot be configured in VLAN database mode, nor stored in the vlan.dat file, nor advertised via VTP. In fact, to configure them, the switch must be in VTP transparent mode. (Also, you should take care to avoid using VLANs 1006–1024 for compatibility with CatOS-based switches.) Both ISL and 802.1Q support extended-range VLANs today. Originally, ISL began life only supporting normal-range VLANs, using only 10 of the 15 bits reserved in the ISL header to identify the VLAN ID. The later-defined 802.1Q used a 12-bit VLAN ID field, thereby allowing support of the extended range. Following that, Cisco changed ISL to use 12 of its reserved 15 bits in the VLAN ID field, thereby supporting the extended range. Table 2-5 summarizes VLAN numbers and provides some additional notes. Table 2-5

Valid VLAN Numbers, Normal and Extended VLAN Number

Normal or Extended?

Can Be Advertised and Pruned by VTP Versions 1 and 2?

Comments

0

Reserved

—

Not available for use

1

Normal

No

On Cisco switches, the default VLAN for all access ports; cannot be deleted or changed

2–1001

Normal

Yes

VLAN Trunking Protocol

Table 2-5

47

Valid VLAN Numbers, Normal and Extended (Continued) VLAN Number

Normal or Extended?

Can Be Advertised and Pruned by VTP Versions 1 and 2?

1002–1005

Normal

No

1006–4094

Extended

No

Comments Defined specifically for use with FDDI and TR translational bridging

Storing VLAN Configuration Catalyst IOS stores VLAN and VTP configuration in one of two places—either in a Flash file called vlan.dat or in the running configuration. (Remember that the term “Catalyst IOS” refers to a switch that uses IOS, not the Catalyst OS, which is often called CatOS.) IOS chooses the storage location in part based on the VTP mode, and in part based on whether the VLANs are normalrange VLANs or extended-range VLANs. Table 2-6 describes what happens based on what configuration mode is used to configure the VLANs, the VTP mode, and the VLAN range. (Note that VTP clients also store the VLAN configuration in vlan.dat, and they do not understand extended range VLANs.) Table 2-6

VLAN Configuration and Storage Function

When in VTP Server Mode

When in VTP Transparent Mode

Normal-range VLANs can be configured from

Both VLAN database and configuration modes

Both VLAN database and configuration modes

Extended-range VLANs can be configured from

Nowhere—cannot be configured

Configuration mode only

VTP and normal-range VLAN configuration commands are stored in

vlan.dat in Flash

Both vlan.dat in Flash and running configuration1

Extended-range VLAN configuration commands stored in

Nowhere—extended range not allowed in VTP server mode

Running configuration only

1When

a switch reloads, if the VTP mode or domain name in the vlan.dat file and the startup-config file differ, the switch uses only the vlan.dat file’s contents for VLAN configuration.

NOTE The configuration characteristics referenced in Table 2-6 do not include the interface configuration command switchport access vlan; it includes the commands that create a VLAN (vlan command) and VTP configuration commands. Of particular interest for those of you stronger with CatOS configuration skills is that when you erase the startup-config file, and reload the Cisco IOS switch, you do not actually erase the

48

Chapter 2: Virtual LANs and VLAN Trunking

normal-range VLAN and VTP configuration information. To erase the VLAN and VTP configuration, you must use the delete flash:vlan.dat exec command. Also note that if multiple switches are in VTP server mode, if you delete vlan.dat on one switch and then reload it, as soon as the switch comes back up and brings up a trunk, it learns the old VLAN database via a VTP update from the other VTP server.

VLAN Trunking: ISL and 802.1Q VLAN trunking allows switches, routers, and even PCs with the appropriate NICs to send traffic for multiple VLANs across a single link. In order to know to which VLAN a frame belongs, the sending switch, router, or PC adds a header to the original Ethernet frame, with that header having a field in which to place the VLAN ID of the associated VLAN. This section describes the protocol details for the two trunking protocols, followed by the details of how to configure trunking.

ISL and 802.1Q Concepts If two devices are to perform trunking, they must agree to use either ISL or 802.1Q, because there are several differences between the two, as summarized in Table 2-7. Table 2-7

Comparing ISL and 802.1Q Feature

ISL

802.1Q

VLANs supported

Normal and extended range1

Normal and extended range

Protocol defined by

Cisco

IEEE

Encapsulates original frame or inserts tag

Encapsulates

Inserts tag

Supports native VLAN

No

Yes

1ISL originally supported only normal-range VLANs, but was later improved to support extended-range VLANs as well.

ISL and 802.1Q differ in how they add a header to the Ethernet frame before sending it over a trunk. ISL adds a new 26-byte header, plus a new trailer (to allow for the new FCS value), encapsulating the original frame. This encapsulating header uses the source address (listed as SA in Figure 2-3) of the device doing the trunking, instead of the source MAC of the original frame. ISL uses a multicast destination address (listed as DA in Figure 2-3) of either 0100.0C00.0000 or 0300.0C00.0000. 802.1Q inserts a 4-byte header, called a tag, into the original frame (right after the Source Address field). The original frame’s addresses are left intact. Normally, an Ethernet controller would expect to find either an Ethernet Type field or 802.3 Length field right after the Source Address field. With an 802.1Q tag, the first 2 bytes after the Address fields holds a registered Ethernet type value of 0x8100, which implies that the frame includes an 802.1Q header. Because 802.1Q does not actually encapsulate the original frame, it is often called frame tagging. Figure 2-3 shows the contents of the headers used by both ISL and 802.1Q.

VLAN Trunking: ISL and 802.1Q

Figure 2-3

49

ISL and 802.1Q Frame Marking Methods ISL Header CRC Encapsulated Ethernet Frame 26 bytes 4 bytes

DA Type User SA LEN AAAA03 HSA VLAN BPDU INDEX RES

VLAN

Dest

Src

Dest

Src

Len/Etype

Etype

Tag

Priority

Data

Len/Etype

BPDU

FCS

Original Frame

Data

FCS

Tagged Frame

VLAN-ID

Finally, the last row from Table 2-7 refers to the native VLAN. 802.1Q does not tag frames sent inside the native VLAN. The native VLAN feature allows a switch to attempt to use 802.1Q trunking on an interface, but if the other device does not support trunking, the traffic for that one native VLAN can still be sent over the link. By default, the native VLAN is VLAN 1.

ISL and 802.1Q Configuration Cisco switches use the Dynamic Trunk Protocol (DTP) to dynamically learn whether the device on the other end of the cable wants to perform trunking and, if so, which trunking protocol to use. DTP learns whether to trunk based on the DTP mode defined for an interface. Cisco switches default to use the DTP desirable mode, which means that the switch initiates sending DTP messages, hoping that the device on the other end of the segment replies with another DTP message. If a reply is received, DTP can detect whether both switches can trunk and, if so, which type of trunking to use. If both switches support both types of trunking, they choose to use ISL. (An upcoming section, “Trunk Configuration Compatibility,” covers the different DTP modes and how they work.) With the DTP mode set to desirable, switches can simply be connected, and they should dynamically form a trunk. You can, however, configure trunking details and verify the results with show commands. Table 2-8 lists some of the key Catalyst IOS commands related to trunking.

50

Chapter 2: Virtual LANs and VLAN Trunking

VLAN Trunking–Related Commands

Table 2-8

Command

Function

switchport | no switchport

Toggle defining whether to treat the interface as a switch interface (switchport) or as a router interface (no switchport)

switchport mode

Sets DTP negotiation parameters

switchport trunk

Sets trunking parameters if the interface is trunking

switchport access

Sets nontrunking-related parameters if the interface is not trunking

show interface trunk

Summary of trunk-related information

show interface type number trunk

Lists trunking details for a particular interface

show interface type number switchport

Lists nontrunking details for a particular interface

Figure 2-4 lists several details regarding Switch1’s trunking configuration and status, as shown in Example 2-5. R1 is not configured to trunk, so Switch1 will fail to negotiate trunking. Switch2 is a Catalyst 3550, which supports both ISL and 802.1Q, so they will negotiate trunking and use ISL. Switch3 and Switch4 are Catalyst 2950s, which support only 802.1Q; as a result, Switch1 negotiates trunking, but picks 802.1Q as the trunking protocol. Figure 2-4

Trunking Configuration Reference for Example 2-5 Not Configured to Trunk Fa0/1

R1

Fa0/12

Defaults to DTP Desirable Does Not Support ISL (2950)

Example 2-5

SW2

Defaults to DTP Desirable Supports ISL or .1Q

SW4

Defaults to DTP Desirable Does Not Support ISL (2950)

Gi0/1

SW1

SW3

Fa0/24

Trunking Configuration and show Command Example–Switch1

! The administrative mode of dynamic desirable (trunking) and negotiate (trunking ! encapsulation) means that Switch1 attempted to negotiate to trunk, but the ! operational mode of static access means that trunking negotiation failed. ! The reference to “operational trunking encapsulation” of native means that ! no tagging occurs.

VLAN Trunking: ISL and 802.1Q

Example 2-5

51

Trunking Configuration and show Command Example–Switch1 (Continued)

Switch1# show int fa 0/1 switchport Name: Fa0/1 Switchport: Enabled Administrative Mode: dynamic desirable Operational Mode: static access Administrative Trunking Encapsulation: negotiate Operational Trunking Encapsulation: native Negotiation of Trunking: On Access Mode VLAN: 1 (default) Trunking Native Mode VLAN: 1 (default) Administrative private-vlan host-association: none Administrative private-vlan mapping: none Operational private-vlan: none Trunking VLANs Enabled: ALL Pruning VLANs Enabled: 2-1001 Protected: false Unknown unicast blocked: disabled Unknown multicast blocked: disabled Voice VLAN: none (Inactive) Appliance trust: none ! Next, the show int gig 0/1 trunk command shows the configured mode ! (desirable), and the current status (N-ISL), meaning negotiated ISL. Note ! that the trunk supports the extended VLAN range as well. Switch1# show int gig 0/1 trunk Port

Mode

Encapsulation

Status

Native vlan

Gi0/1

desirable

n-isl

trunking

1

Port

Vlans allowed on trunk

Gi0/1

1-4094

Port

Vlans allowed and active in management domain

Gi0/1

1,21-22

Port

Vlans in spanning tree forwarding state and not pruned

Gi0/1

1,21-22

! Next, Switch1 lists all three trunks – the segments connecting to the other ! three switches – along with the type of encapsulation. Switch1# show int trunk Port

Mode

Encapsulation

Status

Native vlan

Fa0/12

desirable

n-802.1q

trunking

1

Fa0/24

desirable

n-802.1q

trunking

1

Gi0/1

desirable

n-isl

trunking

1

Port

Vlans allowed on trunk

Fa0/12

1-4094

continues

52

Chapter 2: Virtual LANs and VLAN Trunking

Example 2-5

Trunking Configuration and show Command Example–Switch1 (Continued)

Fa0/24

1-4094

Gi0/1

1-4094

Port

Vlans allowed and active in management domain

Fa0/12

1,21-22

Fa0/24

1,21-22

Gi0/1

1,21-22

Port

Vlans in spanning tree forwarding state and not pruned

Fa0/12

1,21-22

Fa0/24

1,21-22

Gi0/1

1,21-22

Allowed, Active, and Pruned VLANs Although a trunk can support VLANs 1–4094, several mechanisms reduce the actual number of VLANs whose traffic flows over the trunk. First, VLANs can be administratively forbidden from existing over the trunk using the switchport trunk allowed interface subcommand. Also, any allowed VLANs must be configured on the switch before they are considered active on the trunk. Finally, VTP can prune VLANs from the trunk, with the switch simply ceasing to forward frames from that VLAN over the trunk. The show interface trunk command lists the VLANs that fall into each category, as shown in the last command in Example 2-5. The categories are summarized as follows: ■

Allowed VLANs—Each trunk allows all VLANs by default. However, VLANs can be removed or added to the list of allowed VLANs by using the switchport trunk allowed command.

■

Allowed and active—To be active, a VLAN must be in the allowed list for the trunk (based on trunk configuration), and the VLAN must exist in the VLAN configuration on the switch. With PVST+, an STP instance is actively running on this trunk for the VLANs in this list.

■

Active and not pruned—This list is a subset of the “allowed and active” list, with any VTP-pruned VLANs removed.

Trunk Configuration Compatibility In most production networks, switch trunks are configured using the same standard throughout the network. For instance, rather than allow DTP to negotiate trunking,, many engineers configure trunk interfaces to always trunk (switchport mode trunk) and disable DTP on ports that should not trunk. IOS includes several commands that impact whether a particular segment becomes a trunk. Because many enterprises use a typical standard, it is easy to forget the nuances of how the related commands work. This section covers those small details.

VLAN Trunking: ISL and 802.1Q

53

Two IOS configuration commands impact if and when two switches form a trunk. The switchport mode and switchport nonegotiate interface subcommands define whether DTP even attempts to negotiate a trunk, and what rules it uses when the attempt is made. Additionally, the settings on the switch ports on either side of the segment dictate whether a trunk forms or not. Table 2-9 summarizes the trunk configuration options. The first column suggests the configuration on one switch, with the last column listing the configuration options on the other switch that would result in a working trunk between the two switches. Table 2-9

Trunking Configuration Options That Lead to a Working Trunk To Trunk, Other Side Must Be

Configuration Command on One Side1

Short Name

switchport mode trunk

Trunk

Always trunks on this end; sends DTP to help other side choose to trunk

On, desirable, auto

switchport mode trunk; switchport nonegotiate

Nonegotiate

Always trunks on this end; does not send DTP messages (good when other switch is a non-Cisco switch)

On

switchport mode dynamic desirable

Desirable

Sends DTP messages, and trunks if negotiation succeeds

On, desirable, auto

switchport mode dynamic auto

Auto

Replies to DTP messages, and trunks if negotiation succeeds

On, desirable

switchport mode access

Access

Never trunks; sends DTP to help other side reach same conclusion

(Never trunks)

switchport mode access; switchport nonegotiate

Access (with nonegotiate)

Never trunks; does not send DTP messages

(Never trunks)

Meaning

1When the switchport nonegotiate command is not listed in the first column, the default (DTP negotiation is active) is assumed.

NOTE If an interface trunks, then the type of trunking (ISL or 802.1Q) is controlled by the setting on the switchport trunk encapsulation command. This command includes an option for dynamically negotiating the type (using DTP) or configuring one of the two types. See Example 2-5 for a sample of the syntax.

Configuring Trunking on Routers VLAN trunking can be used on routers and hosts as well as on switches. However, routers do not support DTP, so you must manually configure them to support trunking. Additionally, you must manually configure a switch on the other end of the segment to trunk, because the router does not participate in DTP.

54

Chapter 2: Virtual LANs and VLAN Trunking

The majority of router trunking configurations use subinterfaces, with each subinterface being associated with one VLAN. The subinterface number does not have to match the VLAN ID; rather, the encapsulation command sits under each subinterface, with the associated VLAN ID being part of the encapsulation command. Also, because good design calls for one IP subnet per VLAN, if the router wants to forward IP packets between the VLANs, the router needs to have an IP address associated with each trunking subinterface. You can configure 802.1Q native VLANs under a subinterface or under the physical interface on a router. If configured under a subinterface, you use the encapsulation dot1q vlan-id native subcommand, with the inclusion of the native keyword meaning that frames exiting this subinterface should not be tagged. As with other router trunking configurations, the associated IP address would be configured on that same subinterface. Alternately, if not configured on a subinterface, the router assumes that the native VLAN is associated with the physical interface. In this case, the encapsulation command is not needed under the physical interface; the associated IP address, however, would need to be configured under the physical interface. Example 2-6 shows an example configuration for Router1 in Figure 2-1, both for ISL and 802.1Q. In this case, Router1 needs to forward packets between the subnets on VLANs 21 and 22. The first part of the example shows ISL configuration, with no native VLANs, and therefore only a subinterface being used for each VLAN. The second part of the example shows an alternative 802.1Q configuration, using the option of placing the native VLAN (VLAN 21) configuration on the physical interface. Example 2-6

Trunking Configuration on Router1

! Note the subinterface on the fa 0/0 interface, with the encapsulation ! command noting the type of trunking, as well as the VLAN number. The ! subinterface does not have to be the VLAN ID. Also note the IP addresses for ! each interface, allowing Router1 to route between VLANs. interface fastethernet 0/0.1 ip address 10.1.21.1 255.255.255.0 encapsulation isl 21 ! interface fastethernet 0/0.2 ip address 10.1.22.1 255.255.255.0 encapsulation isl 22 ! Next, an alternative 802.1Q configuration is shown. Note that this 802.1Q configuration ! places the IP address ! for VLAN 21 on the physical interface; the router simply associates the ! physical interface with the native VLAN. Alternatively, a subinterface could be ! used, with the encapsulation dot1q 21 native command specifying that the router ! should treat this VLAN as the native VLAN. interface fastethernet 0/0 ip address 10.1.21.1 255.255.255.0 ! interface fastethernet 0/0.2 ip address 10.1.22.1 255.255.255.0 encapsulation dot1q 22

VLAN Trunking: ISL and 802.1Q

55

Note also that the router does not have an explicitly defined allowed VLAN list. However, the allowed VLAN list is implied based on the configured VLANs. For instance, in this example, Router1 allows VLAN 1 (because it cannot be deleted), VLAN 21, and VLAN 22. A show interface trunk command on Switch1 would show only 1, 21, and 22 as the allowed VLANs on FA0/1.

802.1Q-in-Q Tunneling Traditionally, VLANs have not extended beyond the WAN boundary. VLANs in one campus extend to a WAN edge router, but VLAN protocols are not used on the WAN. Today, several emerging alternatives exist for the passage of VLAN traffic across a WAN, including 802.1Q-in-Q, Ethernet over MPLS (EoMPLS), and VLAN MPLS (VMPLS). While these topics are more applicable to the CCIE Service Provider certification, you should at least know the concept of 802.1 Q-in-Q tunneling. Also known as Q-in-Q or Layer 2 protocol tunneling, 802.1Q-in-Q allows an SP to preserve 802.1Q VLAN tags across a WAN service. By doing so, VLANs actually span multiple geographically dispersed sites. Figure 2-5 shows the basic idea.

Figure 2-5

Q-in-Q: Basic Operation SP: Customer1: VLAN 5 Customer2: VLAN 6 Eth. VLAN Header ID 100

C1-SW1 VLANs 100-199

Eth. VLAN Header ID 100

Data

Eth. Header

VLAN ID 5

SP-SW1

C2-SW1 VLANs 100-500 Eth. VLAN Header ID 100

Eth. Header

Data

VLAN ID 100

Data

Data C1-SW2

SP-SW2 VLAN ID 6

VLAN ID 100

Data

Eth. VLAN Header ID 100

C2-SW2

Data

56

Chapter 2: Virtual LANs and VLAN Trunking

The ingress SP switch takes the 802.1Q frame, and then tags each frame entering the interface with an additional 802.1Q header. In this case, all of Customer1’s frames are tagged as VLAN 5 as they pass over the WAN; Customer2’s frames are tagged with VLAN 6. After removing the tag at egress, the customer switch sees the original 802.1Q frame, and can interpret the VLAN ID correctly. The receiving SP switch (SP-SW2 in this case) can keep the various customers’ traffic separate based on the additional VLAN tags. Using Q-in-Q, an SP can offer VLAN services, even when the customers use overlapping VLAN IDs. Customers get more flexibility for network design options, particularly with metro Ethernet services. Plus, CDP and VTP traffic passes transparently over the Q-in-Q service.

Configuring PPPoE Although it might seem out of place in this chapter on VLANs and VLAN trunking, Point-to-Point Protocol over Ethernet (PPPoE) fits best here because it’s an Ethernet encapsulation protocol. PPPoE is widely used for digital subscriber line (DSL) Internet access because the public telephone network uses ATM for its transport protocol; therefore, Ethernet frames must be encapsulated in a protocol supported over both Ethernet and ATM. PPP is the natural choice. The PPP Client feature permits a Cisco IOS router, rather than an endpoint host, to serve as the client in a network. This permits multiple hosts to connect over a single PPPoE connection. In a DSL environment, PPP interface IP addresses are derived from an upstream DHCP server using IP Configuration Protocol (IPCP). Therefore, IP address negotiation must be enabled on the router’s dialer interface. This is done using the ip address negotiated command in the dialer interface configuration. Because of the 8-byte PPP header, the MTU for PPPoE is usually set to 1492 bytes so that the entire encapsulated frame fits within the 1500-byte Ethernet frame. A maximum transmission unit (MTU) mismatch prevents a PPPoE connection from coming up. Checking the MTU setting is a good first step when troubleshooting PPPoE connections. Those familiar with ISDN BRI configuration will recognize the dialer interface configuration and related commands in Example 2-7. The key difference between ISDN BRI configuration and PPPoE is the pppoe-client dial-pool-number command. Configuring an Ethernet edge router for PPPoE Client mode is the focus of this section. This task requires configuring the Ethernet interface (physical or subinterface) and a corresponding dialer interface. The information in this section applies to Cisco IOS Release 12.2(13)T and later, and 12.3 and 12.4 releases.

Configuring PPPoE

57

Figure 2-6 shows the topology. Example 2-7 shows the configuration steps. The first step is to configure the outside Ethernet interface as a PPPoE client and assign it a dialer interface number. The second step is to configure the corresponding dialer interface. Additional steps, including Network Address Translation (NAT) configuration, are also shown.

Figure 2-6

PPPoE Topology for Example 2-7 DSL CPE Fa0/1

Fa0/0 LAN

EdgeRouter Workstations ATM Network

DSL Access Multiplexer

Example 2-7

Access Concentrator

Configuring PPPoE on EdgeRouter

EdgeRouter# conf t EdgeRouter(config)# interface fa0/1 EdgeRouter(config-if)# ip address 192.168.100.1 255.255.255.0 EdgeRouter(config-if)# ip nat inside EdgeRouter(config)# interface fa0/1 EdgeRouter(config-if)# pppoe-client dial-pool-number 1 EdgeRouter(config-if)# exit EdgeRouter(config)# interface dialer1 EdgeRouter(config-if)# mtu 1492 EdgeRouter(config-if)# encapsulation ppp EdgeRouter(config-if)# ip address negotiated EdgeRouter(config-if)# ppp authentication chap !The remaining CHAP commands have been omitted for brevity. EdgeRouter(config-if)# ip nat outside EdgeRouter(config-if)# dialer pool 1

continues

58

Chapter 2: Virtual LANs and VLAN Trunking

Example 2-7

Configuring PPPoE on EdgeRouter (Continued)

EdgeRouter(config-if)# dialer-group 1 EdgeRouter(config-if)# exit EdgeRouter(config)# dialer-list 1 protocol ip permit EdgeRouter(config)# ip nat inside source list 1 interface dialier1 overload EdgeRouter(config)# access-list 1 permit 192.168.100.0 0.0.0.255 EdgeRouter(config)# ip route 0.0.0.0 0.0.0.0 dialer1

You can verify PPPoE connectivity using the command show pppoe session. Cisco IOS includes debug functionality for PPPoE through the debug pppoe [data | errors | events | packets] command.

Foundation Summary

59

Foundation Summary This section lists additional details and facts to round out the coverage of the topics in this chapter. Unlike most of the Cisco Press Exam Certification Guides, this “Foundation Summary” does not repeat information presented in the “Foundation Topics” section of the chapter. Please take the time to read and study the details in the “Foundation Topics” section of the chapter, as well as review items noted with a Key Topic icon. Table 2-10 lists some of the most popular IOS commands related to the topics in this chapter. (The command syntax was copied from the Catalyst 3550 Multilayer Switch Command Reference, 12.1(20)EA2. Note that some switch platforms may have differences in the command syntax.) Table 2-10

Catalyst IOS Commands Related to Chapter 2 Command

Description

show mac address-table [aging-time | count | dynamic | static] [address hw-addr] [interface interface-id] [vlan vlan-id]

Displays the MAC address table; the security option displays information about the restricted or static settings

show interfaces [interface-id | vlan vlan-id] switchport | trunk]

Displays detailed information about an interface operating as an access port or a trunk

show vlan [brief | id vlan-id | name vlan-name | summary]

EXEC command that lists information about VLAN

show vlan [vlan]

Displays VLAN information

show vtp status

Lists VTP configuration and status information

switchport mode {access | dot1q-tunnel | dynamic {auto | desirable} | trunk}

Configuration command setting nontrunking (access), trunking, and dynamic trunking (auto and desirable) parameters

switchport nonegotiate

Interface subcommand that disables DTP messages; interface must be configured as trunk or access port

switchport trunk {allowed vlan vlan-list} | {encapsulation {dot1q | isl | negotiate}} | {native vlan vlan-id} | {pruning vlan vlan-list}

Interface subcommand used to set parameters used when the port is trunking

switchport access vlan vlan-id

Interface subcommand that statically configures the interface as a member of that one VLAN

60

Chapter 2: Virtual LANs and VLAN Trunking

Table 2-11 lists the commands related to VLAN creation—both the VLAN database mode configuration commands (reached with the vlan database privileged mode command) and the normal configuration mode commands. NOTE Some command parameters may not be listed in Table 2-11.

Table 2-11

Catalyst 3550 VLAN Database and Configuration Mode Command List VLAN Database

Configuration

vtp {domain domain-name | password password | pruning | v2-mode | {server | client | transparent}}

vtp {domain domain-name | file filename | interface name | mode {client | server | transparent} | password password | pruning | version number}

vlan vlan-id [backupcrf {enable | disable}] [mtu mtu-size] [name vlan-name] [parent parent-vlan-id] [state {suspend | active}]

vlan vlan-id1

show {current | proposed | difference}

No equivalent

apply | abort | reset

No equivalent

1Creates the VLAN and places the user in VLAN configuration mode, where commands matching the VLAN database

mode options of the vlan command are used to set the same parameters.

Table 2-12

Cisco IOS PPPoE Client Commands Command

Description

pppoe enable

Enables PPPoE operation on an Ethernet interface or subinterface

pppoe-client dial-pool-number number

Configures the outside Ethernet interface on a router for PPPoE operation and ties it to a dialer interface

debug pppoe [data | errors | events | packets]

Enables debugging for PPPoE troubleshooting

Memory Builders The CCIE Routing and Switching written exam, like all Cisco CCIE written exams, covers a fairly broad set of topics. This section provides some basic tools to help you exercise your memory about some of the broader topics covered in this chapter.

Memory Builders

61

Fill In Key Tables from Memory Appendix G, “Key Tables for CCIE Study,” on the CD in the back of this book contains empty sets of some of the key summary tables in each chapter. Print Appendix G, refer to this chapter’s tables in it, and fill in the tables from memory. Refer to Appendix H, “Solutions for Key Tables for CCIE Study,” on the CD to check your answers.

Definitions Next, take a few moments to write down the definitions for the following terms: VLAN, broadcast domain, DTP, VTP pruning, 802.1Q, ISL, native VLAN, encapsulation, private VLAN, promiscuous port, community VLAN, isolated VLAN, 802.1Q-in-Q, Layer 2 protocol tunneling, PPPoE, DSL. Refer to the glossary to check your answers.

Further Reading The topics in this chapter tend to be covered in slightly more detail in CCNP Switching exam preparation books. For more details on these topics, refer to the Cisco Press CCNP preparation books found at www.ciscopress.com/ccnp. Cisco LAN Switching, by Kennedy Clark and Kevin Hamilton, is an excellent reference for LAN-related topics in general, and certainly very useful for CCIE written and lab exam preparation.

Blueprint topics covered in this chapter: This chapter covers the following subtopics from the Cisco CCIE Routing and Switching written exam blueprint. Refer to the full blueprint in Table I-1 in the Introduction for more details on the topics covered in each chapter and their context within the blueprint. ■

Spanning Tree Protocol — 802.1d — 802.1w — 802.1s — Loop Guard — Root Guard — Bridge Protocol Data Unit (BPDU) Guard — Storm Control — Unicast Flooding — STP Port Roles, Failure Propagation, and Loop Guard Operation

■

Troubleshooting Complex Layer 2 Issues

CHAPTER

3

Spanning Tree Protocol Spanning Tree Protocol (STP) is probably one of the most widely known protocols covered on the CCIE Routing and Switching written exam. STP has been around a long time, is used in most every campus network today, and is covered extensively on the CCNP BCMSN exam. This chapter covers a broad range of topics related to STP.

“Do I Know This Already?” Quiz Table 3-1 outlines the major headings in this chapter and the corresponding “Do I Know This Already?” quiz questions. Table 3-1

“Do I Know This Already?” Foundation Topics Section-to-Question Mapping Foundation Topics Section

Questions Covered in This Section

802.1d Spanning Tree Protocol

1–6

Optimizing Spanning Tree

7–9

Protecting STP

10

Troubleshooting Complex Layer 2 Issues

11

Score

Total Score

To best use this pre-chapter assessment, remember to score yourself strictly. You can find the answers in Appendix A, “Answers to the ‘Do I Know This Already?’ Quizzes.” 1.

Assume that a nonroot 802.1d switch has ceased to receive Hello BPDUs. Which STP setting determines how long a nonroot switch waits before trying to choose a new Root Port? a.

Hello timer setting on the Root

b.

Maxage timer setting on the Root

c.

Forward Delay timer setting on the Root

d.

Hello timer setting on the nonroot switch

e.

Maxage timer setting on the nonroot switch

f.

Forward Delay timer setting on the nonroot switch

64

Chapter 3: Spanning Tree Protocol

2.

3.

4.

5.

Assume that a nonroot 802.1d switch receives a Hello BPDU with the TCN flag set. Which STP setting determines how long the nonroot switch waits before timing out CAM entries? a.

Hello timer setting on the Root

b.

Maxage timer setting on the Root

c.

Forward Delay timer setting on the Root

d.

Hello timer setting on the nonroot switch

e.

Maxage timer setting on the nonroot switch

f.

Forward Delay timer setting on the nonroot switch

Assume that nonroot Switch1 (SW1) is blocking on a 802.1Q trunk connected to Switch2 (SW2). Both switches are in the same MST region. SW1 ceases to receive Hellos from SW2. What timers have an impact on how long Switch1 takes to both become the Designated Port on that link and reach forwarding state? a.

Hello timer setting on the Root

b.

Maxage timer setting on the Root

c.

Forward Delay timer on the Root

d.

Hello timer setting on SW1

e.

Maxage timer setting on SW1

f.

Forward Delay timer on SW1

Which of the following statements are true regarding support of multiple spanning trees over an 802.1Q trunk? a.

Only one common spanning tree can be supported.

b.

Cisco PVST+ supports multiple spanning trees if the switches are Cisco switches.

c.

802.1Q supports multiple spanning trees when using IEEE 802.1s MST.

d.

Two PVST+ domains can pass over a region of non-Cisco switches using 802.1Q trunks by encapsulating non-native VLAN Hellos inside the native VLAN Hellos.

When a switch notices a failure, and the failure requires STP convergence, it notifies the Root by sending a TCN BPDU. Which of the following best describes why the notification is needed? a.

To speed STP convergence by having the Root converge quickly.

b.

To allow the Root to keep accurate count of the number of topology changes.

c.

To trigger the process that causes all switches to use a short timer to help flush the CAM.

d.

There is no need for TCN today; it is a holdover from DEC’s STP specification.

“Do I Know This Already?” Quiz

6.

7.

8.

9.

65

Two switches have four parallel Ethernet segments, none of which forms into an EtherChannel. Assuming 802.1d is in use, what is the maximum number of the eight ports (four on each switch) that stabilize into a forwarding state? a.

1

b.

3

c.

4

d.

5

e.

7

Two switches have four Ethernet segments connecting them, with the intention of using an EtherChannel. Port fa 0/1 on one switch is connected to port fa0/1 on the other switch; port fa0/2 is connected to the other switch’s port fa0/2; and so on. An EtherChannel can still form using these four segments, even though some configuration settings do not match on the corresponding ports on each switch. Which settings do not have to match? a.

DTP negotiation settings (auto/desirable/on)

b.

Allowed VLAN list

c.

STP per-VLAN port cost on the ports on a single switch

d.

If 802.1Q, native VLAN

IEEE 802.1w does not use the exact same port states as does 802.1d. Which of the following are valid 802.1w port states? a.

Blocking

b.

Listening

c.

Learning

d.

Forwarding

e.

Disabled

f.

Discarding

What STP tools or protocols supply a “Maxage optimization,” allowing a switch to bypass the wait for Maxage to expire when its Root Port stops receiving Hellos? a.

Loop Guard

b.

UDLD

c.

UplinkFast

d.

BackboneFast

e.

IEEE 802.1w

66

Chapter 3: Spanning Tree Protocol

10.

A trunk between switches lost its physical transmit path in one direction only. Which of the following features protect against the STP problems caused by such an event? a.

Loop Guard

b.

UDLD

c.

UplinkFast

d.

PortFast

802.1d Spanning Tree Protocol

67

Foundation Topics

802.1d Spanning Tree Protocol Although many CCIE candidates already know STP well, the details are easily forgotten. For instance, you can install a campus LAN, possibly turn on a few STP optimizations and security features out of habit, and have a working LAN using STP—without ever really contemplating how STP does what it does. And in a network that makes good use of Layer 3 switching, each STP instance might span only three to four switches, making the STP issues much more manageable— but more forgettable in terms of helping you remember things you need to know for the exam. This chapter reviews the details of IEEE 802.1d STP, and then goes on to related topics—802.1w RSTP, multiple spanning trees, STP optimizations, and STP security features. STP uses messaging between switches to stabilize the network into a logical, loop-free topology. To do so, STP causes some interfaces (popularly called ports when discussing STP) to simply not forward or receive traffic—in other words, the ports are in a blocking state. The remaining ports, in an STP forwarding state, together provide a loop-free path to every Ethernet segment in the network.

Choosing Which Ports Forward: Choosing Root Ports and Designated Ports To determine which ports forward and block, STP follows a three-step process, as listed in Table 3-2. Following the table, each of the three steps is explained in more detail. Table 3-2

Three Major 802.1d STP Process Steps Major Step

Description

Elect the root switch

The switch with the lowest bridge ID wins; the standard bridge ID is 2-byte priority followed by a MAC address unique to that switch.

Determine each switch’s Root Port

The one port on each switch with the least cost path back to the root.

Determine the Designated Port for each segment

When multiple switches connect to the same segment, this is the switch that forwards the least cost Hello onto a segment.

Electing a Root Switch Only one switch can be the root of the spanning tree; to select the root, the switches hold an election. Each switch begins its STP logic by creating and sending an STP Hello bridge protocol

68

Chapter 3: Spanning Tree Protocol

data unit (BPDU) message, claiming to be the root switch. If a switch hears a superior Hello—a Hello with a lower bridge ID—it stops claiming to be root by ceasing to originate and send Hellos. Instead, the switch starts forwarding the superior Hellos received from the superior candidate. Eventually, all switches except the switch with the best bridge ID cease to originate Hellos; that one switch wins the election and becomes the root switch. The original IEEE 802.1d bridge ID held two fields: ■

The 2-byte Priority field, which was designed to be configured on the various switches to affect the results of the STP election process.

■

A 6-byte MAC Address field, which was included as a tiebreaker, because each switch’s bridge ID includes a MAC address value that should be unique to each switch. As a result, some switch must win the root election.

The format of the original 802.1d bridge ID has been redefined. Figure 3-1 shows the original and new format of the bridge IDs. Figure 3-1

IEEE 802.1d STP Bridge ID Formats 2 Bytes

6 Bytes

Priority (0 – 65,535)

System ID (MAC Address)

Original Format Bridge ID

Priority Multiple of 4096

System ID Extension (Typically Holds VLAN ID)

System ID (MAC Address)

4 Bits

12 Bits

6 Bytes

System ID Extension (MAC Address Reduction)

The format was changed mainly due to the advent of multiple spanning trees as supported by Per VLAN Spanning Tree Plus (PVST+) and IEEE 802.1s Multiple Spanning Trees (MST). With the old-style bridge ID format, a switch’s bridge ID for each STP instance (possibly one per VLAN) was identical if the switch used a single MAC address when building the bridge ID. Having multiple STP instances with the same bridge ID was confusing, so vendors such as Cisco Systems used a different Ethernet BIA for each VLAN when creating the old-style bridge IDs. This provided a different bridge ID per VLAN, but it consumed a large number of reserved BIAs in each switch. The System ID Extension allows a network to use multiple instances of STP, even one per VLAN, but without the need to consume a separate BIA on each switch for each STP instance. The System ID Extension field allows the VLAN ID to be placed into what was formerly the last 12 bits of the

802.1d Spanning Tree Protocol

69

Priority field. A switch can use a single MAC address to build bridge IDs, and with the VLAN number in the System ID Extension field still have a unique bridge ID in each VLAN. The use of the System ID Extension field is also called MAC address reduction, because of the need for many fewer reserved MAC addresses on each switch. Determining the Root Port Once the root is elected, the rest of the switches now need to determine their Root Port (RP). The process proceeds as described in the following list: 1.

The root creates and sends a Hello every Hello timer (2 seconds default).

2.

Each switch that receives a Hello forwards the Hello after updating the following fields in the Hello: the cost, the forwarding switch’s bridge ID, forwarder’s port priority, and forwarder’s port number.

3.

Switches do not forward Hellos out ports that stabilize into a blocking state.

4.

Of all the ports in which a switch receives Hellos, the port with the least calculated cost to the root is the RP.

A switch must examine the cost value in each Hello, plus the switch’s STP port costs, in order to determine its least cost path to reach the root. To do so, the switch adds the cost listed in the Hello message to the switch’s port cost of the port on which the Hello was received. For example, Figure 3-2 shows the loop network design and details several STP cost circulations. Figure 3-2

Calculating STP Costs to Determine RPs Loop Design – All Port Costs 19 Unless Shown Root SW1

DP

Cost 19 RP

DP Cost 1

SW2 DP

Hello Cost 0

Fa0/1 Cost 100 BL

SW3

Adding my incoming cost setting fields for forwarder’s bridge ID, port priority, and port number.

Hello Cost 0

Hello Cost 19 Cost out fa0/1 = 0 + 100 = 100; cost out fa0/4 = 38 + 19 = 57! RP Cost 19 RP Fa0/4 Cost 19

DP

SW4 Hello Cost 38

In Figure 3-2, SW1 happened to become root, and is originating Hellos of cost 0. SW3 receives two Hellos, one with cost 0 and one with cost 38. However, SW3 must then calculate its cost to reach the root, which is the advertised cost (0 and 38, respectively) plus SW3’s port costs (100 and 19, respectively). As a result, although SW3 has a direct link to SW1, the calculated cost is lower

70

Chapter 3: Spanning Tree Protocol

out interface fa0/4 (cost 57) than it is out interface fa0/1 (cost 100), so SW3 chooses its fa0/4 interface as its RP. NOTE Many people think of STP costs as being associated with a segment; however, the cost is actually associated with interfaces. Good design practices dictate using the same STP cost on each end of a point-to-point Ethernet segment, but the values can be different.

While the costs shown in Figure 3-2 might seem a bit contrived, the same result would happen with default port costs if the link from SW1 to SW3 were Fast Ethernet (default cost 19), and the other links were Gigabit Ethernet (default cost 4). Table 3-3 lists the default port costs according to IEEE 802.1d. Note that the IEEE updated 802.1d in the late 1990s, changing the suggested default port costs. Table 3-3

Default Port Costs According to IEEE 802.1d Speed of Ethernet

Original IEEE Cost

Revised IEEE Cost

10 Mbps

100

100

100 Mbps

10

19

1 Gbps

1

4

10 Gbps

1

2

When a switch receives multiple Hellos with equal calculated cost, it uses the following tiebreakers: 1.

Pick the lowest value of the forwarding switch’s bridge ID.

2.

Use the lowest port priority of the neighboring switch. The neighboring switch added its own port priority to the Hello before forwarding it.

3.

Use the lowest internal port number (of the forwarding switch) as listed inside the received Hellos.

Note that if the first tiebreaker in this list fails to produce an RP, this switch must have multiple links to the same neighboring switch. The last two tiebreakers simply help decide which of the multiple parallel links to use. Determining the Designated Port A converged STP topology results in only one switch forwarding Hellos onto each LAN segment. The switch that forwards onto a LAN segment is called the designated switch for that segment, and the port that it uses to forward frames onto that segment is called the Designated Port (DP).

802.1d Spanning Tree Protocol

71

To win the right to be the DP, a switch must send the Hello with the lowest advertised cost onto the segment. For instance, consider the segment between SW3 and SW4 in Figure 3-2 before the DP has been determined on that segment. SW3 would get Hellos directly from SW1, compute its cost to the root over that path, and then forward the Hello out its fa 0/4 interface to SW4, with cost 100. Similarly, SW4 will forward a Hello with cost 38, as shown in Figure 3-2. SW4’s fa 0/3 port becomes the DP due to its lower advertised cost. Only the DP forwards Hellos onto a LAN segment as well. In the same example, SW4 keeps sending the cost-38 Hellos out the port, but SW3 stops sending its inferior Hellos. When the cost is a tie, STP uses the same tiebreakers to choose the DP as when choosing an RP: lowest forwarder’s bridge ID, lowest forwarder’s port priority, and lowest forwarder’s port number.

Converging to a New STP Topology STP logic monitors the normal ongoing Hello process when the network topology is stable; when the Hello process changes, STP then needs to react and converge to a new STP topology. When STP has a stable topology, the following occurs: 1.

The root switch generates a Hello regularly based on the Hello timer.

2.

Each nonroot switch regularly (based on the Hello timer) receives a copy of the root’s Hello on its RP.

3.

Each switch updates and forwards the Hello out its Designated Ports.

4.

For each blocking port, the switch regularly receives a copy of the Hello from the DP on that segment. (The switches do not forward Hellos out blocking interfaces.)

When some deviation from these events occurs, STP knows that the topology has changed and that convergence needs to take place. For instance, one simple case might be that the root switch loses power; the rest of the switches will not hear any Hello messages, and after the Maxage timer expires (default 10 times Hello, or 20 seconds), the switches elect a new root based on the logic described earlier in this chapter. For a more subtle example, consider Figure 3-3, which shows the same loop network as in Figure 3-2. In this case, however, the link from SW1 to SW2 has just failed.

72

Chapter 3: Spanning Tree Protocol

Figure 3-3

Reacting to Loss of Link Between SW1 and SW2 Loop Design – All Port Costs 19 Unless Shown 1

MAC 0200.1111.1111

Root SW1

R1

My RP failed. I am receiving no other Hellos. I must be the root now! Disabled

Disabled

Cost 1

1

SW2 Fa0/4 Hello Root = Sw2 Cost 0

Hello Root = Sw1 Cost 0

Fa0/1 Cost 100

SW1’s bridge ID is better. So I’m sending the superior Hello on this segment. I am now DP!

2

Fa0/2 Fa0/3

SW3 4

Fa0/4 Cost 19

Hello Root = Sw1 Cost 100

SW4 Hello Root = Sw2 Cost 19

3

The following list describes some of the key steps from Figure 3-3: 1.

SW2 ceases to receive Hellos on its RP.

2.

Because SW2 is not receiving Hellos over any other path, it begins a new root election by claiming to be root and flooding Hellos out every port.

3.

SW4 notices that the latest Hello implies a new root switch, but SW4 ends up with the same RP (for now). SW4 forwards the Hello out toward SW3 after updating the appropriate fields in the Hello.

4.

SW3 receives the Hello from SW4, but it is inferior to the one SW3 receives from SW1. So, SW3 becomes the DP on the segment between itself and SW4, and starts forwarding the superior Hello on that port.

Remember, SW1 had won the earlier election; as of Steps 3 and 4, the Hellos from SW1 and SW2 are competing, and the one claiming SW1 as root will again win. The rest of the process results with SW2’s fa0/4 as DP, SW4’s fa 0/3 as RP, SW4’s fa 0/2 as DP, and SW2’s fa 0/4 as RP. Topology Change Notification and Updating the CAM When STP reconvergence occurs, some Content Addressable Memory (CAM) entries might be invalid (CAM is the Cisco term for what’s more generically called the MAC address table, switching table, or bridging table on a switch). For instance, before the link failure shown in Figure 3-3, SW3’s CAM might have had an entry for 0200.1111.1111 (Router1’s MAC address) pointing out fa0/4 to SW4. (Remember, at the beginning of the scenario described in Figure 3-3,

802.1d Spanning Tree Protocol

73

SW3 was blocking on its fa0/1 interface back to SW1.) When the link between SW1 and SW2 failed, SW3 would need to change its CAM entry for 0200.1111.111 to point out port fa0/1. To update the CAMs, two things need to occur: ■

All switches need to be notified to time out their CAM entries.

■

Each switch needs to use a short timer, equivalent to the Forward Delay timer (default 15 seconds), to time out the CAM entries.

Because some switches might not directly notice a change in the STP topology, any switch that detects a change in the STP topology has a responsibility to notify the rest of the switches. To do so, a switch simply notifies the root switch in the form of a Topology Change Notification (TCN) BPDU. The TCN goes up the tree to the root. After that, the root notifies all the rest of the switches. The process runs as follows: 1.

A switch experiencing the STP port state change sends a TCN BPDU out its Root Port; it repeats this message every Hello time until it is acknowledged.

2.

The next switch receiving that TCN BPDU sends back an acknowledgment via its next forwarded Hello BPDU by marking the Topology Change Acknowledgment (TCA) bit in the Hello.

3.

The switch that was the DP on the segment in the first two steps repeats the first two steps, sending a TCN BPDU out its Root Port, and awaiting acknowledgment from the DP on that segment.

By each successive switch repeating Steps 1 and 2, eventually the root receives a TCN BPDU. Once received, the root sets the TC flag on the next several Hellos, which are forwarded to all switches in the network, notifying them that a change has occurred. A switch receiving a Hello BPDU with the TC flag set uses the short (Forward Delay time) timer to time out entries in the CAM. Transitioning from Blocking to Forwarding When STP reconverges to a new, stable topology, some ports that were blocking might have been designated as DP or RP, so these ports need to be in a forwarding state. However, the transition from blocking to forwarding state cannot be made immediately without the risk of causing loops. To transition to forwarding state but also prevent temporary loops, a switch first puts a formerly blocking port into listening state, and then into learning state, with each state lasting for the length of time defined by the forward delay timer (by default, 15 seconds). Table 3-4 summarizes the key points about all of the 802.1d STP port states.

74

Chapter 3: Spanning Tree Protocol

IEEE 802.1d Spanning Tree Interface States

Table 3-4

State

Forwards Data Frames?

Learn Source MACs of Received Frames?

Transitory or Stable State?

Blocking

No

No

Stable

Listening

No

No

Transitory

Learning

No

Yes

Transitory

Forwarding

Yes

Yes

Stable

Disabled

No

No

Stable

In summary, when STP logic senses a change in the topology, it converges, possibly picking different ports as RP, DP, or neither. Any switch changing its RPs or DPs sends a TCN BPDU to the root at this point. For the ports newly designated as RP or DP, 802.1d STP first uses the listening and learning states before reaching the forwarding state. (The transition from forwarding to blocking can be made immediately.)

Per-VLAN Spanning Tree and STP over Trunks If only one instance of STP was used for a switched network with redundant links but with multiple VLANs, several ports would be in a blocking state, unused under stable conditions. The redundant links would essentially be used for backup purposes. The Cisco Per VLAN Spanning Tree Plus (PVST+) feature creates an STP instance for each VLAN. By tuning STP configuration per VLAN, each STP instance can use a different root switch and have different interfaces block. As a result, the traffic load can be balanced across the available links. For instance, in the common building design with distribution and access links in Figure 3-4, focus on the left side of the figure. In this case, the access layer switches block on different ports on VLANs 1 and 2, with different root switches. Figure 3-4

Operation of PVST+ for Better Load Balancing ISL

3550

3550

Root VLAN1

Root VLAN2 .1Q

.1Q

.1Q

.1Q

FWD VLAN1

FWD VLAN2 FWD VLAN2

2950

FWD VLAN1

2950

With different root switches and with default port costs, the access layer switches end up sending VLAN1 traffic over one uplink and VLAN2 traffic over another uplink.

802.1d Spanning Tree Protocol

75

Using 802.1Q with STP requires some extra thought as to how it works. 802.1Q does not support PVST+ natively; however, Cisco switches do support PVST+ over 802.1Q trunks. So, with all Cisco switches, and PVST+ (which is enabled by default), PVST+ works fine. When using 802.1Q with non-Cisco switches, the switches must follow the IEEE standard completely, so the trunks support only a Common Spanning Tree (CST). With standard 802.1Q, only one instance of STP runs only over VLAN 1, and that one STP topology is used for all VLANs. Although using only one STP instance reduces the STP messaging overhead, it does not allow load balancing by using multiple STP instances, as was shown with PVST+ in Figure 3-4. When building networks using a mix of Cisco and non-Cisco switches, along with 802.1Q trunking, you can still take advantage of multiple STP instances in the Cisco portion of the network. Figure 3-5 shows two general options in which two CST regions of non-Cisco switches connect to two regions of Cisco PVST+ supporting switches. Figure 3-5

Combining Standard IEEE 802.1Q and CST with PVST+ CST Region 1 Non-Cisco CST Region 2

.1Q

Non-Cisco .1Q

Non-Cisco

Non-Cisco

STP Only Over Native VLAN

Non-Cisco

PVST + Region 3550 .1Q

PVST + Region

Non-native VLAN STP BPDUs trunked, sent to 0100.0CCC.CCCD

3550

ISL .1Q

Non-Cisco

3550

.1Q

.1Q

.1Q .1Q

2950

3550

ISL

.1Q

2950

2950

.1Q

2950

The left side of Figure 3-5 shows an example in which the CST region is not used for transit between multiple PVST+ regions. In this case, none of the PVST+ per-VLAN STP information needs to pass over the CST region. The PVST+ region maps the single CST instance to each of the PVST+ STP instances.

76

Chapter 3: Spanning Tree Protocol

The rest of Figure 3-5 shows two PVST+ regions, separated by a single CST region (CST Region 2). In this case, the PVST+ per-VLAN STP information needs to pass through the CST region. To do so, PVST+ treats the CST region as a single link and tunnels the PVST+ BPDUs across the CST region. The tunnel is created by sending the BPDUs using a multicast destination MAC of 0100.0CCC.CCCD, with the BPDUs being VLAN tagged with the correct VLAN ID. As a result, the non-Cisco switches forward the BPDUs as a multicast, and do not interpret the frames as BPDUs. When a forwarded BPDU reaches the first Cisco PVST+ switch in the other PVST+ region, the switch, listening for multicasts to 0100.0CCC.CCCD, reads and interprets the BPDU. NOTE 802.1Q, along with 802.1s Multiple-instance Spanning Tree (MST), allows 802.1Q trunks for support multiple STP instances. MST is covered later in this chapter.

STP Configuration and Analysis Example 3-1, based on Figure 3-6, shows some of the basic STP configuration and show commands. Take care to note that many of the upcoming commands allow the parameters to be set for all VLANs by omitting the VLAN parameter, or set per VLAN by including a VLAN parameter. Example 3-1 begins with SW1 coincidentally becoming the root switch. After that, SW2 is configured to become root, and SW3 changes its Root Port as a result of a configured port cost in VLAN 1. Figure 3-6

Network Used with Example 3-1 Core Design Fa0/2

Fa0/1

SW1 Fa0/3

SW2 Fa0/4

Fa0/3

Fa0/2

Fa0/4

Fa0/1

Fa0/1

SW3 Example 3-1

Fa0/2 Fa0/4

Fa0/3

SW4

STP Basic Configuration and show Commands

! First, note the Root ID column lists the root’s bridge ID as two parts, ! first the priority, followed by the MAC address of the root. The root cost of ! 0 implies that SW1 (where the command is executed) is the root. SW1# sh spanning-tree root Vlan

Root ID

Root

Hello Max Fwd

Cost

Time

Age Dly

---------------- -------------------- --------- ----- --- --VLAN0001

32769 000a.b7dc.b780

0

2

20

15

VLAN0011

32779 000a.b7dc.b780

0

2

20

15

VLAN0012

32780 000a.b7dc.b780

0

2

20

15

VLAN0021

32789 000a.b7dc.b780

0

2

20

15

VLAN0022

32790 000a.b7dc.b780

0

2

20

15

Root Port ------------

802.1d Spanning Tree Protocol

Example 3-1

77

STP Basic Configuration and show Commands (Continued)

! The next command confirms that SW1 believes that it is the root of VLAN 1. SW1# sh spanning-tree vlan 1 root detail Root ID

Priority

32769

Address

000a.b7dc.b780

This bridge is the root Hello Time

2 sec

Max Age 20 sec

Forward Delay 15 sec

! Next, SW2 is configured with a lower (better) priority than SW1, ! so it becomes the root. Note that because SW2 is defaulting to use ! the system ID extension, the actual priority must be configured as a ! multiple of 4096. SW2# conf t Enter configuration commands, one per line.

End with CNTL/Z.

SW2(config)# spanning-tree vlan 1 priority ?

bridge priority in increments of 4096

SW2(config)# spanning-tree vlan 1 priority 28672 SW2(config)# ^Z SW2# sh spanning-tree vlan 1 root detail VLAN0001 Root ID

Priority

28673

Address

0011.92b0.f500

This bridge is the root Hello Time

2 sec

Max Age 20 sec

Forward Delay 15 sec

! The System ID Extension field of the bridge ID is implied next. The output ! does not separate the 4-bit Priority field from the System ID field. The output ! actually shows the first 2 bytes of the bridge ID, in decimal. For VLAN1, ! the priority is 28,763, which is the configured 28,672 plus the VLAN ID, ! because the VLAN ID value is used in the System ID field in order to implement ! the MAC address reduction feature. The other VLANs have a base priority ! of 32768, plus the VLAN ID - for example, VLAN11 has priority 32779, ! (priority 32,768 plus VLAN 11), VLAN12 has 32780, and so on. SW2# sh spanning-tree root priority VLAN0001

28673

VLAN0011

32779

VLAN0012

32780

VLAN0021

32789

VLAN0022

32790

! Below, SW3 shows a Root Port of Fa 0/2, with cost 19. SW3 gets Hellos ! directly from the root (SW2) with cost 0, and adds its default cost (19). ! This next command also details the breakdown of the priority and system ID. SW3# sh spanning-tree vlan 1 VLAN0001 Spanning tree enabled protocol ieee Root ID

Priority

28673

Address

0011.92b0.f500

Cost

19

continues

78

Chapter 3: Spanning Tree Protocol

Example 3-1

STP Basic Configuration and show Commands (Continued) Port

2 (FastEthernet0/2)

Hello Time Bridge ID

2 sec

Max Age 20 sec

Priority

32769

Address

000e.837b.3100

Hello Time

2 sec

Forward Delay 15 sec

(priority 32768 sys-id-ext 1) Max Age 20 sec

Forward Delay 15 sec

Aging Time 300 Interface

Role Sts Cost

Prio.Nbr Type

---------------- ---- --- --------- -------- -------------------------------Fa0/1

Altn BLK 19

128.1

P2p

Fa0/2

Root FWD 19

128.2

P2p

Fa0/4

Desg FWD 19

128.4

P2p

Fa0/13

Desg FWD 100

128.13

Shr

! Above, the port state of BLK and FWD for each port is shown, as well as the ! Root port and the Designated Ports. ! Below, Switch3’s VLAN 1 port cost is changed on its Root Port (fa0/2), ! causing SW3 to reconverge, and pick a new RP. SW3# conf t Enter configuration commands, one per line.

End with CNTL/Z.

SW3(config)# int fa 0/2 SW3(config-if)# spanning-tree vlan 1 cost 100 SW3(config-if)# ^Z ! The next command was done immediately after changing the port cost on ! SW3. Note the state listed as “LIS,” meaning listen. STP has already ! chosen fa 0/1 as the new RP, but it must now transition through listening ! and learning states. SW3# sh spanning-tree vlan 1 VLAN0001 Spanning tree enabled protocol ieee Root ID

Priority

28673

Address

0011.92b0.f500

Cost

38

Port

1 (FastEthernet0/1)

Hello Time Bridge ID

2 sec

Max Age 20 sec

Priority

32769

Address

000e.837b.3100

Hello Time

2 sec

Forward Delay 15 sec

(priority 32768 sys-id-ext 1) Max Age 20 sec

Forward Delay 15 sec

Aging Time 15 Interface

Role Sts Cost

Prio.Nbr Type

---------------- ---- --- --------- -------- -------------------------------Fa0/1

Root LIS 19

128.1

P2p

Fa0/2

Altn BLK 100

128.2

P2p

Fa0/4

Desg FWD 19

128.4

P2p

Fa0/13

Desg FWD 100

128.13

Shr

Optimizing Spanning Tree

79

The preceding example shows one way to configure the priority to a lower value to become the root. Optionally, the spanning-tree vlan vlan-id root {primary | secondary} [diameter diameter] command could be used. This command causes the switch to set the priority lower. The optional diameter parameter causes this command to lower the Hello, Forward Delay, and Maxage timers. (This command does not get placed into the configuration, but rather it acts as a macro, being expanded into the commands to set priority and the timers.) NOTE When using the primary option, the spanning-tree vlan command sets the priority to 24,576 if the current root has a priority larger than 24,576. If the current root’s priority is 24,576 or less, this command sets this switch’s priority to 4096 less than the current root. With the secondary keyword, this switch’s priority is set to 28,672. Also note that this logic applies to when the configuration command is executed; it does not dynamically change the priority if another switch later advertises a better priority.

Optimizing Spanning Tree Left to default settings, IEEE 802.1d STP works, but convergence might take up to a minute or more for the entire network. For instance, when the root fails, a switch must wait on the 20-second Maxage timer to expire. Then, newly forwarding ports spend 15 seconds each in listening and learning states, which makes convergence take 50 seconds for that one switch. Over the years, Cisco added features to its STP code, and later the IEEE made improvements as well. This section covers the key optimizations to STP.

PortFast, UplinkFast, and BackboneFast The Cisco-proprietary PortFast, UplinkFast, and BackboneFast features each solve specific STP problems. Table 3-5 summarizes when each is most useful, and the short version of how they improve convergence time. Table 3-5

PortFast, UplinkFast, and BackboneFast Feature

Requirements for Use

How Convergence Is Optimized

PortFast

Used on access ports that are not connected to other switches or hubs

Immediately puts the port into forwarding state once the port is physically working

UplinkFast

Used on access layer switches that have multiple uplinks to distribution/core switches

Immediately replaces a lost RP with an alternate RP, immediately forwards on the RP, and triggers updates of all switches’ CAMs

BackboneFast

Used to detect indirect link failures, typically in the network core

Avoids waiting for Maxage to expire when its RP ceases to receive Hellos; does so by querying the switch attached to its RP

80

Chapter 3: Spanning Tree Protocol

PortFast PortFast optimizes convergence by simply ignoring listening and learning states on ports. In effect, convergence happens instantly on ports with PortFast enabled. Of course, if another switch is connected to a port on which PortFast is enabled, loops may occur. So, PortFast is intended for access links attached to single end-user devices. To be safe, you should also enable the BPDU Guard and Root Guard features when using PortFast, as covered later in this chapter. UplinkFast UplinkFast optimizes convergence when an uplink fails on an access layer switch. For good STP design, access layer switches should not become root or become transit switches. (A transit switch is a switch that forwards frames between other switches.) Figure 3-7 shows the actions taken when UplinkFast is enabled on access layer switch SW3, and then when the Root Port fails. Figure 3-7

UplinkFast Operations Access Design Root Switch Priority 8192

Non-Root Switch Priority (Default) 32768

SW1

SW2

Fa0/3

2

Dst. MAC Src. MAC 0100.0CCD.CDCD 0200.3333.3333

RP Link Failure Occurs

4 1

Enable Uplink Fast: Sets priority to 49,152 Sets port costs to 3000 Tracks alternate RPs

RP: FWD

Fa0/1

Fa0/2

Sends multicasts for each source MAC for local MACs

Alt. RP: BLK

3

Makes Fa0/2 the RP Moves to forwarding state immediately

SW3

MAC 0200.3333.3333

R3

Upon enabling UplinkFast globally in a switch, the switch takes three actions: ■

Increases the root priority to 49,152

■

Sets the post costs to 3000

■

Tracks alternate RPs, which are ports in which root Hellos are being received

As a result of these steps, SW3 can become root if necessary, but it is unlikely to do so given the large root priority value. Also, the very large costs on each link make the switch unlikely to be used as a transit switch. When the RP port does fail, SW3 can fail over to an alternate uplink as the new RP and forward immediately.

Optimizing Spanning Tree

81

The final step in Uplink Fast logic causes the switches to time-out the correct entries in their CAMs, but it does not use the TCN process. Instead, the access switch finds all the MAC addresses of local devices and sends one multicast frame with each local MAC address as the source MAC— causing all the other switches to update their CAMs. The access switch also clears out the rest of the entries in its own CAM. BackboneFast BackboneFast optimizes convergence for any generalized topological case, improving convergence when an indirect failure occurs. When some direct failures occur (for instance, a switch’s RP interface fails), the switch does not have to wait for Maxage to expire. However, when another switch’s direct link fails, resulting in lost Hellos for other switches, the downstream switches indirectly learn of the failure because they cease to receive Hellos. Any time a switch learns of an STP failure indirectly, the switch must wait for Maxage to expire before trying to change the STP topology. BackboneFast simply causes switches that indirectly learn of a potential STP failure to ask their upstream neighbors if they know about the failure. To do so, when the first Hello goes missing, a BackboneFast switch sends a Root Link Query (RLQ) BPDU out the port in which the missing Hello should have arrived. The RLQ asks the neighboring switch if that neighboring switch is still receiving Hellos from the root. If that neighboring switch had a direct link failure, it can tell the original switch (via another RLQ) that this path to the root is lost. Once known, the switch experiencing the indirect link failure can go ahead and converge without waiting for Maxage to expire. NOTE All switches must have BackboneFast configured for it to work correctly.

PortFast, UplinkFast, and BackboneFast Configuration Configuration of these three STP optimizing tools is relatively easy, as summarized in Table 3-6. Table 3-6

PortFast, UplinkFast, and BackboneFast Configuration Feature

Configuration Command

PortFast

spanning-tree portfast (interface subcommand) spanning-tree portfast default (global)

UplinkFast

spanning-tree uplinkfast [max-update-rate rate] (global)

BackboneFast

spanning-tree backbonefast (global)

82

Chapter 3: Spanning Tree Protocol

PortChannels When a network design includes multiple parallel segments between the same pair of switches, one switch ends up in a forwarding state on all the links, but the other switch blocks all but one of the ports of those parallel segments. As a result, only one of the links can be used at any point in time. Using Fast EtherChannel (FEC) (using FastE segments) and Gigabit EtherChannel (GEC) (using GigE segments) allows the combined links to be treated as one link from an STP perspective, so that all the parallel physical segments are used. (When configuring a Cisco switch, a group of segments comprising an FEC or GEC is called a PortChannel.) Most campus designs today use a minimum of two segments per trunk, in a PortChannel, for better availability. That way, as long as at least one of the links in the EtherChannel is up, the STP path cannot fail, and no STP convergence is required. Load Balancing Across PortChannels When a switch decides to forward a frame out a PortChannel, the switch must also decide which physical link to use to send each frame. To use the multiple links, Cisco switches load balance the traffic over the links in an EtherChannel based on the switch’s global load-balancing configuration. Load-balancing methods differ depending on the model of switch and software revision. Generally, load balancing is based on the contents of the Layer 2, 3, and/or 4 headers. If load balancing is based on only one header field in the frame, a bitmap of the low-order bits is used; if more than one header field is used, an XOR of the low-order bits is used. For the best balancing effect, the header fields on which balancing is based need to vary among the mix of frames sent over the PortChannel. For instance, for a Layer 2 PortChannel connected to an access layer switch, most of the traffic going from the access layer switch to the distribution layer switch is probably going from clients to the default router. So most of the frames have different source MAC addresses, but the same destination MAC address. For packets coming back from a distribution switch toward the access layer switch, many of the frames might have a source address of that same router, with differing destination MAC addresses. So, you could balance based on source MAC at the access layer switch, and based on destination MAC at the distribution layer switch—or balance based on both fields on both switches. The goal is simply to use a balancing method for which the fields in the frames vary. The port-channel load-balance type command sets the type of load balancing. The type options include using source and destination MAC, IP addresses, and TCP and UDP ports—either a single field or both the source and destination.

Optimizing Spanning Tree

83

PortChannel Discovery and Configuration You can explicitly configure interfaces to be in a PortChannel by using the channel-group number mode on interface subcommand. You would simply put the same command under each of the physical interfaces inside the PortChannel, using the same PortChannel number. You can also use dynamic protocols to allow neighboring switches to figure out which ports should be part of the same PortChannel. Those protocols are the Cisco-proprietary Port Aggregation Protocol (PAgP) and the IEEE 802.1AD Link Aggregation Control Protocol (LACP). To dynamically form a PortChannel using PAgP, you still use the channel-group command, with a mode of auto or desirable. To use LACP to dynamically create a PortChannel, use a mode of active or passive. Table 3-7 lists and describes the modes and their meanings. PAgP and LACP Configuration Settings and Recommendations

Table 3-7

PAgP Setting

LACP 802.1AD Setting

Action

on

on

Disables PAgP or LACP, and forces the port into the PortChannel

off

off

Disables PAgP or LACP, and prevents the port from being part of a PortChannel

auto

passive

Uses PAgP or LACP, but waits on other side to send first PAgP or LACP message

desirable

active

Uses PAgP or LACP, and initiates the negotiation

NOTE Using auto (PAgP) or passive (LACP) on both switches prevents a PortChannel from forming dynamically. Cisco recommends the use of desirable mode (PAgP) or active mode (LACP) on ports that you intend to be part of a PortChannel. When PAgP or LACP negotiate to form a PortChannel, the messages include the exchange of some key configuration information. As you might imagine, they exchange a system ID to determine which ports connect to the same two switches. The two switches then exchange other information about the candidate links for a PortChannel; several items must be identical on the links for them to be dynamically added to the PortChannel, as follows: ■

Same speed and duplex settings.

■

If not trunking, same access VLAN.

■

If trunking, same trunk type, allowed VLANs, and native VLAN.

■

On a single switch, each port in a PortChannel must have the same STP cost per VLAN on all links in the PortChannel.

■

No ports can have SPAN configured.

84

Chapter 3: Spanning Tree Protocol

When PAgP or LACP completes the process, a new PortChannel interface exists, and is used as if it were a single port for STP purposes, with balancing taking place based on the global loadbalancing method configured on each switch.

Rapid Spanning Tree Protocol IEEE 802.1w Rapid Spanning Tree Protocol (RSTP) enhances the 802.1d standard with one goal in mind: improving STP convergence. To do so, RSTP defines new variations on BPDUs between switches, new port states, and new port roles, all with the capability to operate backwardly compatible with 802.1d switches. The key components of speeding convergence with 802.1w are as follows: ■

Waiting for only three missed Hellos on an RP before reacting (versus ten missed Hellos via the Maxage timer with 802.1d)

■

New processes that allow transition from the disabled state (replaces the blocking state in 802.1d) to learning state, bypassing the concept of an 802.1d listening state

■

Standardization of features like Cisco PortFast, UplinkFast, and BackboneFast

■

An additional feature to allow a backup DP when a switch has multiple ports connected to the same shared LAN segment

To support these new processes, RSTP uses the same familiar Hello BPDUs, using some previously undefined bits to create the new features. For instance, RSTP defines a Hello message option for the same purpose as the Cisco proprietary RLQ used by the Cisco BackboneFast feature. RSTP takes advantage of a switched network topology by categorizing ports, using a different link type to describe each. RSTP takes advantage of the fact that STP logic can be simplified in some cases, based on what is attached to each port, thereby allowing faster convergence. Table 3-8 lists the three RSTP link types. Table 3-8

RSTP Link Types Link Type

Description

Point-to-point

Connects a switch to one other switch; Cisco switches treat FDX links in which Hellos are received as point-to-point links.

Shared

Connects a switch to a hub; the important factor is that switches are reachable off that port.

Edge

Connects a switch to a single end-user device.

Optimizing Spanning Tree

85

In most modern LAN designs with no shared hubs, all links would be either the point-to-point (a link between two switches) or edge link type. RSTP knows that link-type edge means the port is cabled to one device, and the device is not a switch. So, RSTP treats edge links with the same logic as Cisco PortFast—in fact, the same spanning-tree portfast command defines a port as link-type edge to RSTP. In other words, RSTP puts edge links into forwarding state immediately. RSTP takes advantage of point-to-point links (which by definition connect a switch to another switch) by asking the other switch about its status. For instance, if one switch fails to receive its periodic Hello on a point-to-point link, it will query the neighbor. The neighbor will reply, stating whether it also lost its path to the root. It is the same logic as BackboneFast, but using IEEE standard messages to achieve the same goal. RSTP also redefines the port states used with 802.1d, in part because the listening state is no longer needed. Table 3-9 compares the port states defined by each protocol. Table 3-9

RSTP and STP Port States Administrative State

STP State (802.1d)

RSTP State (802.1w)

Disabled

Disabled

Discarding

Enabled

Blocking

Discarding

Enabled

Listening

Discarding

Enabled

Learning

Learning

Enabled

Forwarding

Forwarding

In RSTP, a discarding state means that the port does not forward frames, receive frames, or learn source MAC addresses, regardless of whether the port was shut down, failed, or simply does not have a reason to forward. Once RSTP decides to transition from discarding to forwarding state (for example, a newly selected RP), it goes immediately to the learning state. From that point on, the process continues just as it does with 802.1d. RSTP no longer needs the listening state because of its active querying to neighbors, which guarantees no loops during convergence. RSTP uses the term port role to refer to whether a port acts as an RP or a DP. RSTP uses the RP and DP port roles just as 802.1d does; however, RSTP adds several other roles, as listed in Table 3-10.

86

Chapter 3: Spanning Tree Protocol

Table 3-10

RSTP and STP Port Roles RSTP Role

Definition

Root Port

Same as 802.1d Root Port.

Designated Port

Same as 802.1d Designated Port.

Alternate Port

Same as the Alternate Port concept in UplinkFast; an alternate Root Port.

Backup Port

A port that is attached to the same link-type shared link as another port on the same switch, but the other port is the DP for that segment. The Backup Port is ready to take over if the DP fails.

The Alternate Port concept is like the UplinkFast concept—it offers protection against the loss of a switch’s RP by keeping track of the Alternate Ports with a path to the root. The concept and general operation is identical to UplinkFast, although RSTP might converge more quickly via its active messaging between switches. The Backup Port role has no equivalent with Cisco-proprietary features; it simply provides protection against losing the DP attached to a shared link when the switch has another physical port attached to the same shared LAN. You can enable RSTP in a Cisco switch by using the spanning-tree mode rapid-pvst global command. Alternatively, you can simply enable 802.1s MST, which by definition uses 802.1w RSTP. Rapid Per-VLAN Spanning Tree Plus (RPVST+) RPVST+ is a combination of PVST+ and RSTP. This combination provides the subsecond convergence of RSTP with the advantages of PVST+ described in the previous section. Thus, RPVST+ and PVST+ share the same characteristics such as convergence time, Hello behavior, the election process, port states, and so forth. RPVST+ is compatible with MSTP and PVST+. Configuring RPVST+ is straightforward. In global configuration mode, issue the spanning-tree mode rapid-pvst command. Then, optionally, on an interface (VLAN, physical, or PortChannel), configure the spanning-tree link-type point-to-point command. This configures the port for fast changeover to the forwarding state. See the “Further Reading” section for a source of more information on RPVST+.

Optimizing Spanning Tree

87

Multiple Spanning Trees: IEEE 802.1s IEEE 802.1s Multiple Spanning Trees (MST), sometimes referred to as Multiple Instance STP (MISTP) or Multiple STP (MSTP), defines a way to use multiple instances of STP in a network that uses 802.1Q trunking. The following are some of the main benefits of 802.1s: ■

Like PVST+, it allows the tuning of STP parameters so that while some ports block for one VLAN, the same port can forward in another VLAN.

■

Always uses 802.1w RSTP, for faster convergence.

■

Does not require an STP instance for each VLAN; rather, the best designs use one STP instance per redundant path.

If the network consists of all MST-capable switches, MST is relatively simple to understand. A group of switches that together uses MST is called an MST region; to create an MST region, the switches need to be configured as follows: 1.

Globally enable MST, and enter MST configuration mode by using the spanning-tree mode mst command.

2.

From MST configuration mode, create an MST region name (up to 32 characters) by using the name subcommand.

3.

From MST configuration mode, create an MST revision number by using the revision command.

4.

From MST configuration mode, map VLANs to an MST STP instance by using the instance command.

The key to MST configuration is to configure the same parameters on all the switches in the region. For instance, if you match VLANs 1–4 to MST instance 1 on one switch, and VLANs 5–8 to MST instance 1 on another switch, the two switches will not consider themselves to be in the same MST region, even though their region names and revision numbers are identical. For example, in Figure 3-8, an MST region has been defined, along with connections to nonMST switches. Focusing on the left side of the figure, inside the MST region, you really need only two instances of STP—one each for roughly half of the VLANs. With two instances, the access layer switches will forward on their links to SW1 for one set of VLANs using one MST instance, and forward on their links to SW2 for the other set of VLANs using the second MST instance.

88

Chapter 3: Spanning Tree Protocol

Figure 3-8

MST Operations Non-MST Region

Non-Cisco

Non-MST Region

Non-Cisco

.1Q

.1Q

.1Q

Non-Cisco

Non-Cisco

IST Topology

MST Region 3550 Root Instance1

3550

ISL

.1Q

.1Q

.1Q FWD Instance1

2950

Root Instance2

MST appears as a single switch to the outside world

.1Q

FWD Instance2

FWD Instance1

3550

FWD Instance2

2950

One of the key benefits of MST versus PVST+ is that it requires only one MST instance for a group of VLANs. If this MST region had hundreds of VLANs, and used PVST+, hundreds of sets of STP messages would be used. With MST, only one set of STP messages is needed for each MST instance. When connecting an MST region to a non-MST region or to a different MST region, MST makes the entire MST region appear to be a single switch, as shown on the right side of Figure 3-8. An MST region can guarantee loop-free behavior inside the MST region. To prevent loops over the CST links connecting the MST region to a non-MST region, MST participates in an STP instance with the switches outside the MST region. This additional STP instance is called the Internal Spanning Tree (IST). When participating in STP with the external switches, the MST region is made to appear as if it is a single switch; the right side of Figure 3-8 depicts the STP view of the left side of the figure, as seen by the external switches.

Protecting STP This section covers four switch configuration tools that protect STP from different types of problems or attacks, depending on whether a port is a trunk or an access port.

Protecting STP

89

Root Guard and BPDU Guard: Protecting Access Ports Network designers probably do not intend for end users to connect a switch to an access port that is intended for attaching end-user devices. However, it happens—for instance, someone just may need a few more ports in the meeting room down the hall, so they figure they could just plug a small, cheap switch into the wall socket. The STP topology can be changed based on one of these unexpected and undesired switches being added to the network. For instance, this newly added and unexpected switch might have the lowest bridge ID and become the root. To prevent such problems, BPDU Guard and Root Guard can be enabled on these access ports to monitor for incoming BPDUs—BPDUs that should not enter those ports, because they are intended for single end-user devices. Both features can be used together. Their base operations are as follows: ■

BPDU Guard—Enabled per port; error disables the port upon receipt of any BPDU.

■

Root Guard—Enabled per port; ignores any received superior BPDUs to prevent a switch connected to this port from becoming root. Upon receipt of superior BPDUs, this switch puts the port in a loop-inconsistent state, ceasing forwarding and receiving frames until the superior BPDUs cease.

With BPDU Guard, the port does not recover from the err-disabled state unless additional configuration is added. You can tell the switch to change from err-disabled state back to an up state after a certain amount of time. With Root Guard, the port recovers when the undesired superior BPDUs are no longer received.

UDLD and Loop Guard: Protecting Trunks Both UniDirectional Link Detection (UDLD) and Loop Guard protect a switch trunk port from causing loops. Both features prevent switch ports from errantly moving from a blocking to a forwarding state when a unidirectional link exists in the network. Unidirectional links are simply links for which one of the two transmission paths on the link has failed, but not both. This can happen as a result of miscabling, cutting one fiber cable, unplugging one fiber, GBIC problems, or other reasons. Although UDLD was developed for fiber links because of the unidirectional nature of fiber optic cabling—and therefore the much greater likelihood of a unidirectional link failure in a fiber cable—this feature also supports copper links. Because STP monitors incoming BPDUs to know when to reconverge the network, adjacent switches on a unidirectional link could both become forwarding, causing a loop, as shown in Figure 3-9.

90

Chapter 3: Spanning Tree Protocol

Figure 3-9

STP Problems with Unidirectional Links One Trunk, Two Fiber Cables DP FWD Tx

SW1 RP FWD

2 Hello Cost 19

1

3

Non-DP BLK

No more Hellos. I must be the DP. Let me transition to forwarding!

SW2

Rx

RP FWD

DP FWD

Root

SW3

DP FWD

Figure 3-9 shows the fiber link between SW1 and SW2 with both cables. SW2 starts in a blocking state, but as a result of the failure on SW1’s transmit path, SW2 ceases to hear Hellos from SW1. SW2 then transitions to forwarding state, and now all trunks on all switches are forwarding. Even with the failure of SW1’s transmit cable, frames will now loop counter-clockwise in the network. UDLD uses two modes to attack the unidirectional link problem. As described next, both modes, along with Loop Guard, solve the STP problem shown in Figure 3-9: ■

UDLD—Uses Layer 2 messaging to decide when a switch can no longer receive frames from a neighbor. The switch whose transmit interface did not fail is placed into an err-disabled state.

■

UDLD aggressive mode—Attempts to reconnect with the other switch (eight times) after realizing no messages have been received. If the other switch does not reply to the repeated additional messages, both sides become err-disabled.

■

Loop Guard—When normal BPDUs are no longer received, the port does not go through normal STP convergence, but rather falls into an STP loop-inconsistent state.

In all cases, the formerly blocking port that would now cause a loop is prevented from migrating to a forwarding state. With both types of UDLD, the switch can be configured to automatically transition out of err-disabled state. With Loop Guard, the switch automatically puts the port back into its former STP state when the original Hellos are received again.

Troubleshooting Complex Layer 2 Issues

91

Troubleshooting Complex Layer 2 Issues Troubleshooting is one of the most challenging aspects of CCIE study. The truth is, we can’t teach you to troubleshoot in the pages of a book; only time and experience bring strong troubleshooting skills. We can, however, provide you with two things that are indispensable in learning to troubleshoot effectively and efficiently: process and tools. The focus of this section is to provide you with a set of Cisco IOS–based tools, beyond the more common ones that you already know, as well as some guidance on the troubleshooting process for Layer 2 issues that you might encounter. In the CCIE Routing and Switching lab exam, you will encounter an array of troubleshooting situations that require you to have mastered fast, efficient, and thorough troubleshooting skills. In the written exam, you’ll need a different set of skills—mainly, the knowledge of troubleshooting techniques that are specific to Cisco routers and switches, and the ability to interpret the output of various show commands and possibly debug output. You can also expect to be given an example along with a problem statement. You will need to quickly narrow the question down to possible solutions, and then pinpoint the final solution. This requires a different set of skills than what the lab exam requires, but spending time on fundamentals as you prepare for the qualification exam will provide a good foundation for the lab exam environment. As in all CCIE exams, you should expect that the easiest or most direct ways to a solution might be unavailable to you. In troubleshooting, perhaps the easiest way to the source of most problems is through the show run command or variations on it. Therefore, we’ll institute a simple “no show run” rule in this section that will force you to use your knowledge of more in-depth troubleshooting commands in the Cisco IOS portion of this section. In addition, you can expect that the issues you’ll face in this part of the written exam will need more than one command or step to isolate and resolve.

Layer 2 Troubleshooting Process From the standpoint of troubleshooting techniques, two basic stack-based approaches come into play depending on what type of issue you’re facing. The first of these is the climb-the-stack approach, where you begin at Layer 1 and work your way up until you find the problem. Alternatively, you can start at Layer 7 and work your way down; however, in the context of the CCIE Routing and Switching exams, the climb-the-stack approach generally makes more sense. The second approach is often referred to as the divide-and-conquer method. With this technique, you start in the middle of the stack (usually where you see the problem) and work your way down or up the stack from there until you find the problem. In the interest of time, which is paramount in an exam environment, the divide-and-conquer approach usually provides the best results. Because this section deals with Layer 2 issues, it starts at the bottom and works up.

92

Chapter 3: Spanning Tree Protocol

Some lower-level issues that might affect Layer 2 connectivity include the following: ■

Cabling—Check the physical soundness of the cable as well as the use of a correctly pinned cable. If the switch does not support Automatic Medium-Dependent Interface Crossover (Auto-MDIX), the correct choice of either crossover or straight-through cable must be made.

■

Speed or duplex mismatch—Most Cisco devices will correctly sense speed and duplex when both sides of the link are set to Auto, but a mismatch will cause the line protocol on the link to stay down.

■

Device physical interface—It is possible for a physical port to break.

Layer 2 Protocol Troubleshooting and Commands In addition to the protocol-specific troubleshooting commands that you have learned so far, this section addresses commands that can help you isolate problems through a solid understanding of the information they present. We will use a variety of examples of command output to illustrate the key parameters you should understand. Troubleshooting Using Basic Interface Statistics The command show interfaces is a good place to start troubleshooting interface issues. It will tell you whether the interface has a physical connection and whether it was able to form a logical connection. The link duplex and bandwidth are shown, along with errors and collisions. Example 3-2 shows output from this command, with important statistics highlighted. Example 3-2

Troubleshooting with the show interface Command

sw4# show int fa0/21 !Shows a physical and logical connection FastEthernet0/21 is up, line protocol is up (connected) Hardware is Fast Ethernet, address is 001b.d4b3.8717 (bia 001b.d4b3.8717) MTU 1500 bytes, BW 100000 Kbit, DLY 100 usec, reliability 255/255, txload 1/255, rxload 1/255 Encapsulation ARPA, loopback not set !Interval for Layer 2 keepalives. This should match on both sides of the link Keep alive set (10 sec) !Negotiated or configured speed and duplex Full-duplex, 100Mb/s, media type is 10/100BaseTX input flow-control is off, output flow-control is unsupported ARP type: ARPA, ARP Timeout 04:00:00 Last input 00:00:01, output 00:00:08, output hang never Last clearing of "show interface" counters never Input queue: 0/75/0/0 (size/max/drops/flushes); Total output drops: 0 Queueing strategy: fifo Output queue: 0/40 (size/max)

Troubleshooting Complex Layer 2 Issues

Example 3-2

93

Troubleshooting with the show interface Command (Continued) 5 minute input rate 0 bits/sec, 0 packets/sec 5 minute output rate 0 bits/sec, 0 packets/sec 16206564 packets input, 1124307496 bytes, 0 no buffer Received 14953512 broadcasts (7428112 multicasts) !CRC errors, runts, frames, collisions or late collisions !may indicated a duplex mismatch 0 runts, 0 giants, 0 throttles 0 input errors, 0 CRC, 0 frame, 0 overrun, 0 ignored 0 watchdog, 7428112 multicast, 0 pause input 0 input packets with dribble condition detected 2296477 packets output, 228824856 bytes, 0 underruns 0 output errors, 0 collisions, 1 interface resets 0 babbles, 0 late collision, 0 deferred 0 lost carrier, 0 no carrier, 0 PAUSE output 0 output buffer failures, 0 output buffers swapped out

If an interface shows as up/up, you know that a physical and logical connection has been made, and you can move on up the stack in troubleshooting. If it shows as up/down, you have some Layer 2 troubleshooting to do. An interface status of err-disable could be caused by many different problems, some of which are discussed later in this chapter. Common causes include a security violation or detection of a unidirectional link. Occasionally, a duplex mismatch will cause this state. Chapter 1, “Ethernet Basics,” showed examples of a duplex mismatch, but the topic is important enough to include here. Duplex mismatch might be caused by hard-coding one side of the link to full duplex but leaving the other side to autonegotiate duplex. A 10/100 interface will default to half duplex if the other side is 10/100 and does not negotiate. It could also be caused by an incorrect manual configuration on both sides of the link. A duplex mismatch usually does not bring the link down; it just creates suboptimal performance. You would suspect a duplex mismatch if you saw collisions on a full-duplex link, because a fullduplex link should never have collisions. A link that is half duplex on both sides will show some interface errors. But more than about 1 percent to 2 percent of the total traffic is cause for a second look. Watch for the following types of errors: ■

Runts—Runts are frames smaller than 64 bytes.

■

CRC errors—The frame’s cyclic redundancy checksum value does not match the one calculated by the switch or router.

■

Frames—Frame errors have a CRC error and contain a noninteger number of octets.

94

Chapter 3: Spanning Tree Protocol

■

Alignment—Alignment errors have a CRC error and an odd number of octets.

■

Collisions—Look for collisions on a full-duplex interface, or excessive collisions on a halfduplex interface.

■

Late collisions on a half-duplex interface—A late collision occurs after the first 64 bytes of a frame.

Another command to display helpful interface statistics is show controllers, shown in Example 3-3. The very long output from this command is another place to find the number of frames with bad Frame Checks, CRC errors, collisions, and late collisions. In addition, it tells you the size breakdown of frames received and transmitted. A preponderance of one-size frames on an interface that is performing poorly can be a clue to the application sending the frames. Another useful source of information is the interface autonegotiation status and the speed/duplex capabilities of it and its neighbor, shown at the bottom of Example 3-3. Example 3-3

Troubleshooting with the show controllers Command

R1# show controllers fastEthernet 0/0 Interface FastEthernet0/0 Hardware is MV96340 HWIDB: 46F92948, INSTANCE: 46F939F0, FASTSEND: 4374CB14, MCI_INDEX: 0 Aggregate MIB Counters ---------------------Rx Good Bytes: 27658728

Rx Good Frames: 398637

Rx Bad Bytes: 0

Rx Bad Frames: 0

Rx Broadcast Frames: 185810

Rx Multicast Frames: 181353

Tx Good Bytes: 3869662

Tx Good Frames: 36667

Tx Broadcast Frames: 0

Tx Multicast Frames: 5684

Rx+Tx Min-64B Frames: 412313

Rx+Tx 65-127B Frames: 12658

Rx+Tx 128-255B Frames: 0

Rx+Tx 256-511B Frames: 10333

Rx+Tx 512-1023B Frames: 0

Rx+Tx 1024-MaxB Frames: 0

Rx Unrecog MAC Ctrl Frames: 0 Rx Good FC Frames: 0

Rx Bad FC Frames: 0

Rx Undersize Frames: 0

Rx Fragment Frames: 0

Rx Oversize Frames: 0

Rx Jabber Frames: 0

Rx MAC Errors: 0

Rx Bad CRCs: 0

Tx Collisions: 0

Tx Late Collisions: 0

![output omitted] AUTONEG_EN PHY Status (0x01): AUTONEG_DONE LINK_UP Auto-Negotiation Advertisement (0x04): 100FD 100HD 10FD 10HD Link Partner Ability (0x05): 100FD 100HD 10FD 10HD !output omitted

Troubleshooting Complex Layer 2 Issues

95

Troubleshooting Spanning Tree Protocol Spanning-tree issues are possible in a network that has not been properly configured. Previous sections of this chapter discussed ways to secure STP. One common STP problem is a change in the root bridge. If the root bridge is not deterministically configured, a change in the root can cause a flood of BPDUs and affect network connectivity. To lessen the chance of this, use Rapid STP and all the tools necessary to secure the root and user ports. Example 3-1 showed commands to check the root bridge and other STP parameters, including the following: show spanning-tree root [p p riority ] root detail ] show spanning-tree vlan number [r

Keep in mind that when BPDU Guard is enabled, a port is error-disabled if it receives a BPDU. You can check this with the command show interfaces status err-disabled. In addition, switching loops can result if BPDU Guard is enabled on a trunk port or an interface has a duplex mismatch. One symptom of a loop is flapping MAC addresses. A port with Root Guard or Loop Guard configured is put in an inconsistent state if it receives a superior BPDU. You can check this with the command show spanning-tree inconsistentports. Whether an interface is error-disabled or put into an inconsistent state, the port is effectively shut down to user traffic. Troubleshooting Trunking Trunks that fail to form may result from several causes. With an 802.1Q trunk, a native VLAN mismatch is usually the first thing troubleshooters look at. You should additionally check the Dynamic Trunking Protocol (DTP) negotiation mode of each side of the trunk. Table 2-9 in Chapter 2, “Virtual LANs and VLAN Trunking,” lists the combinations of DTP configurations that will lead to successful trunking. A VLAN Trunking Protocol (VTP) domain mismatch has been known to prevent trunk formation, even in switches that are in VTP Transparent mode. The switch’s logging output will help you greatly. This is shown in Example 3-4, along with some commands that will help you troubleshoot trunking problems. In Example 3-4, two switches are configured with 802.1Q native VLANs 10 and 99, and DTP mode desirable. Both are VTP transparent and have different VTP domain names. Some output irrelevant to the example is omitted. Example 3-4

Troubleshooting Trunking

!These errors messages were logged by the switch %CDP-4-NATIVE_VLAN_MISMATCH: Native VLAN mismatch discovered on FastEthernet1/0/21 (10), with sw4 FastEthernet0/21 (99) %SPANTREE-2-RECV_PVID_ERR: Received BPDU with inconsistent peer vlan id 99 on FastEthernet1/0/21 VLAN10 %DTP-5-DOMAINMISMATCH: Unable to perform trunk negotiation on port Fa1/0/21 because of VTP domain mismatch. ! !This command shows that the port is configured to trunk (Administrative Mode) but is not performing as a trunk

continues

96

Chapter 3: Spanning Tree Protocol

Example 3-4

Troubleshooting Trunking (Continued)

!(Operational Mode) SW2# show int fa 1/0/1 switchport Name: Fa1/0/1 Switchport: Enabled Administrative Mode: dynamic desirable Operational Mode: static access Administrative Trunking Encapsulation: negotiate Operational Trunking Encapsulation: native Negotiation of Trunking: On Access Mode VLAN: 1 (default) Trunking Native Mode VLAN: 10 (NATIVE_10) Administrative Native VLAN tagging: enabled !output omitted ! Trunking VLANs Enabled: 3,99 ! !The port is shown as inconsistent due to native VLAN mismatch sw4# show spanning-tree inconsistentports Name

Interface

Inconsistency

------------ ------------------------ -----------------VLAN0099

FastEthernet0/21

Port VLAN ID Mismatch

Number of inconsistent ports (segments) in the system : 1 ! !Once the errors are corrected, the interface shows as a trunk sw4# show interfaces trunk Port

Mode

Encapsulation

Status

Native vlan

Fa0/21

desirable

802.1q

trunking

99

!output omitted

If your trunks are connected and operating, but user connectivity is not working, check the VLANs allowed on each trunk. Make sure that the allowed VLANs match on each side of the trunk, and that the users’ VLAN is on the allowed list (assuming it should be). Either look at the interface configuration or use the show interfaces switchport command shown in Example 3-4 to find that information. Troubleshooting VTP If you choose to use anything other than VTP Transparent mode in your network, you should be aware of the ways to break it. VTP will fail to negotiate a neighbor status if the following items do not match: ■

VTP version

■

VTP domain

■

VTP password

Troubleshooting Complex Layer 2 Issues

97

In addition, recall that VTP runs over trunk links only, so you must have an operational trunk before it will update. To prevent your VLAN database from being altered when adding a switch to the VTP domain, follow these steps: Step 1

Change the VTP mode to Transparent, which will reset the configuration revision number to 0.

Step 2

Delete the vlan.dat file from the switch’s flash.

Step 3

Reboot the switch.

Step 4

Configure the appropriate VTP parameters.

Step 5

Configure trunking.

Step 6

Connect the switch to the network.

The first part of Example 3-5 shows a VTP client with a password that doesn’t match its neighbor (note the error message). The switch does not show an IP address in the last line because it has not been able to negotiate a VTP relationship with its neighbor. In the second part of the example, the configuration has been corrected. Now the neighbor’s IP address is listed as the VTP updater. Example 3-5

Troubleshooting VTP

!Wrong password is configured sw4# show vtp status VTP Version

: running VTP1 (VTP2 capable)

Configuration Revision

: 0

Maximum VLANs supported locally : 1005 Number of existing VLANs

: 5

VTP Operating Mode

: Client

VTP Domain Name

: CCIE

VTP Pruning Mode

: Disabled

VTP V2 Mode

: Disabled

VTP Traps Generation

: Disabled

MD5 digest

: 0xA1 0x7C 0xE8 0x7E 0x4C 0xF5 0xE3 0xC8

*** MD5 digest checksum mismatch on trunk: Fa0/23 *** *** MD5 digest checksum mismatch on trunk: Fa0/24 *** Configuration last modified by 0.0.0.0 at 7-24-09 03:12:27 ! !Command output after the misconfiguration was corrected sw4# show vtp status VTP Version

: running VTP2

Configuration Revision

: 5

Maximum VLANs supported locally : 1005 Number of existing VLANs

: 9

VTP Operating Mode

: Client

continues

98

Chapter 3: Spanning Tree Protocol

Example 3-5

Troubleshooting VTP (Continued)

VTP Domain Name

: CCIE

VTP Pruning Mode

: Disabled

VTP V2 Mode

: Enabled

VTP Traps Generation

: Disabled

MD5 digest

: 0xDD 0x6C 0x64 0xF5 0xD2 0xFE 0x9B 0x62

Configuration last modified by 192.168.250.254 at 7-24-09 11:02:43

Troubleshooting EtherChannels Table 3-7 listed the LACP and PAgP settings. If your EtherChannel is not coming up, check these settings. If you are using LACP, at least one side of each link must be set to “active.” If you are using PagP, at least one side of the link must be set to “desirable.” If you are not using a channel negotiation protocol, make sure that both sides of the links are set to “on.” Remember that the following rules apply to all ports within an EtherChannel: ■

Speed and duplex must match.

■

Interface type—access, trunk, or routed—must match.

■

Trunk configuration—encapsulation, allowed VLANs, native VLAN, and DTP mode—must match.

■

If a Layer 2 EtherChannel is not a trunk, all ports must be assigned to the same VLAN.

■

No port in the EtherChannel can be a Switched Port Analyzer (SPAN) port.

■

On a Layer 3 EtherChannel, the IP address must be on the PortChannel interface, not a physical interface.

To troubleshoot an EtherChannel problem, check all the parameters in the preceding list. Example 3-6 shows some commands to verify the logical and physical port configuration for an EtherChannel. QoS configuration must match and must be configured on the physical ports, not the logical one. Example 3-6

Troubleshooting EtherChannels

!The show etherchannel summary command gives an overview of the !channels configured, whether they are Layer 2 or Layer 3, the !interfaces assigned to each, and the protocol used if any L3SW4# show etherchannel summary Flags:

D - down

P - bundled in port-channel

I - stand-alone s - suspended H - Hot-standby (LACP only)

Troubleshooting Complex Layer 2 Issues

Example 3-6

Troubleshooting EtherChannels (Continued) R - Layer3

S - Layer2

U - in use

f - failed to allocate aggregator

M - not in use, minimum links not met u - unsuitable for bundling w - waiting to be aggregated d - default port Number of channel-groups in use: 3 Number of aggregators: Group

Port-channel

3

Protocol

Ports

------+-------------+---------+------------------------------------------14

Po14(SU)

24

Po24(RU)

34

Po34(RU)

LACP PAgP

Fa0/3(P) Fa0/7(P)

Fa0/8(P)

Fa0/1(P)

Fa0/2(P)

Fa0/9(P)

Fa0/10(P)

! !The show interface etherchannel command lets you verify that the interface is configured with the right channel group and !protocol settings ! L3SW3# show int fa0/1 etherchannel Port state

= Up

Mstr In-Bndl

Channel group = 34

Mode = On

Gcchange = -

Port-channel

= Po34

GC

Pseudo port-channel = Po34

Port index

= 0

Load = 0x00

=

-

Protocol =

PAgP

Age of the port in the current state: 1d:07h:28m:19s ! !The show interface portchannel command produces output similar !to a physical interface. It allows you to verify the ports !assigned to the channel and the type of QoS used L3SW3# show int port-channel 23 Port-channel23 is up, line protocol is up (connected) Hardware is EtherChannel, address is 001f.2721.8643 (bia 001f.2721.8643) Internet address is 10.1.253.13/30 MTU 1500 bytes, BW 200000 Kbit, DLY 100 usec, reliability 255/255, txload 1/255, rxload 1/255 Encapsulation ARPA, loopback not set Keepalive set (10 sec) Full-duplex, 100Mb/s, link type is auto, media type is unknown input flow-control is off, output flow-control is unsupported Members in this channel: Fa0/3 Fa0/4 ARP type: ARPA, ARP Timeout 04:00:00 Last input 00:00:02, output 00:00:00, output hang never Last clearing of "show interface" counters never Input queue: 0/75/0/0 (size/max/drops/flushes); Total output drops: 0 Queueing strategy: fifo

99

100

Chapter 3: Spanning Tree Protocol

Approaches to Resolving Layer 2 Issues In this final section of the chapter, Table 3-11 presents some generalized types of Layer 2 issues and ways of approaching them, including the relevant Cisco IOS commands. Table 3-11

Layer 2 Troubleshooting Approach and Commands Problem

Approach

Helpful IOS Commands

Lack of reachability to devices in the same VLAN

Eliminate Layer 1 issues with show interface commands.

show interface show vlan

Verify that the VLAN exists on the switch.

Intermittent reachability to devices in the same VLAN

No connectivity between switches

show interface switchport

Verify that the interface is assigned to the correct VLAN.

traceroute mac source-mac

Verify that the VLAN is allowed on the trunk.

show interface trunk

Check for excessive interface traffic.

show interface

Check for unidirectional links.

show spanning-tree

Check for spanning-tree problems such as BPDU floods or flapping MAC addresses.

show spanning-tree root

Check for interfaces shut down.

show interfaces status err-disabled

Verify that trunk links and EtherChannels are active. Verify that BPDU Guard is not enabled on a trunk interface.

destination-mac

show mac-address-table

show interface trunk show etherchannel summary show spanning-tree detail

Poor performance across a link

Check for a duplex mismatch.

show interface

In summary, when troubleshooting Layer 2 issues check for interface physical problems or configuration mismatches. Verify that STP is working as expected. If you are using VTP, make sure that it is configured properly on each switch. For trunking problems, check native VLAN and DTP configuration. When troubleshooting port channels, verify that the interface parameters are the same on both sides.

Foundation Summary

101

Foundation Summary This section lists additional details and facts to round out the coverage of the topics in this chapter. Unlike most of the Cisco Press Exam Certification Guides, this “Foundation Summary” does not repeat information presented in the “Foundation Topics” section of the chapter. Please take the time to read and study the details in the “Foundation Topics” section of the chapter, as well as review items noted with a Key Topic icon. Table 3-12 lists the protocols mentioned in this chapter and their respective standards documents. Table 3-12

Protocols and Standards for Chapter 3 Name

Standards Body

RSTP

IEEE 802.1w

MST

IEEE 802.1s

STP

IEEE 802.1d

LACP

IEEE 802.1AD

Dot1Q trunking

IEEE 802.1Q

PVST+

Cisco

RPVST+

Cisco

PagP

Cisco

Table 3-13 lists the three key timers that impact STP convergence. Table 3-13

IEEE 802.1d STP Timers Timer

Default

Purpose

Hello

2 sec

Interval at which the root sends Hellos

Forward Delay

15 sec

Time that switch leaves a port in listening state and learning state; also used as the short CAM timeout timer

Maxage

20 sec

Time without hearing a Hello before believing that the root has failed

Table 3-14 lists some of the key IOS commands related to the topics in this chapter. (The command syntax for switch commands was taken from the Catalyst 3560 Switch Command Reference, 12.2(44)SE).

102

Chapter 3: Spanning Tree Protocol

Table 3-14

Command Reference for Chapter 3 Command

Description

spanning-tree mode {mst | pvst | rapid-pvst}

Global config command that sets the STP mode

[no] spanning-tree vlan x

Enables or disables STP inside a particular VLAN when using PVST+

spanning-tree vlan vlan-id {forwardtime seconds | hello-time seconds | max-age seconds | priority priority | {root {primary | secondary} [diameter net-diameter [hello-time seconds]]}}

Global config command to set a variety of STP parameters

spanning-tree vlan x cost y

Interface subcommand used to set interface costs, per VLAN

spanning-tree vlan x port-priority y

Interface subcommand used to set port priority, per VLAN

channel-group channel-group-number mode {auto [non-silent] | desirable [non-silent] | on | active | passive}

Interface subcommand that places the interface into a port channel, and sets the negotiation parameters

channel-protocol {lacp | pagp}

Interface subcommand to define which protocol to use for EtherChannel negotiation

interface port-channel port-channelnumber

Global command that allows configuration of parameters for the EtherChannel

spanning-tree portfast

Interface subcommand that enables PortFast on the interface

spanning-tree bpduguard {enable | disable

Interface command that enables or disables BPDU Guard on the interface

spanning-tree uplinkfast

Global command that enables UplinkFast

spanning-tree backbonefast

Global command that enables BackboneFast

spanning-tree mst instance-id priority priority

Global command used to set the priority of an MST instance

spanning-tree mst configuration

Global command that puts user in MST configuration mode

show spanning-tree root | brief | summary

EXEC command to show various details about STP operation

show spanning-tree uplinkfast | backbonefast

EXEC command to show various details about UplinkFast and BackboneFast

show interface

Displays Layer 1 and 2 information about an interface

show interface trunk

Displays the interface trunk configuration

show etherchannel [summary]

Lists EtherChannels configured and their status

show interface switchport

Displays the interface trunking and VLAN configuration

Memory Builders

Table 3-14

103

Command Reference for Chapter 3 (Continued) Command

Description

show vtp status

Displays the VTP configuration

show controllers

Displays physical interface characteristics as well as traffic and error types

Memory Builders The CCIE Routing and Switching written exam, like all Cisco CCIE written exams, covers a fairly broad set of topics. This section provides some basic tools to help you exercise your memory about some of the broader topics covered in this chapter.

Fill in Key Tables from Memory Appendix G, “Key Tables for CCIE Study,” on the CD in the back of this book contains empty sets of some of the key summary tables in each chapter. Print Appendix G, refer to this chapter’s tables in it, and fill in the tables from memory. Refer to Appendix H, “Solutions for Key Tables for CCIE Study,” on the CD to check your answers.

Definitions Next, take a few moments to write down the definitions for the following terms: CST, STP, MST, RSTP, Hello timer, Maxage timer, Forward Delay timer, blocking state, forwarding state, listening state, learning state, disabled state, alternate state, discarding state, backup state, Root Port, Designated Port, superior BPDU, PVST+, RPVST+, UplinkFast, BackboneFast, PortFast, Root Guard, BPDU Guard, UDLD, Loop Guard, LACP, PAgP Refer to the glossary to check your answers.

Further Reading The topics in this chapter tend to be covered in slightly more detail in CCNP Switching exam preparation books. For more details on these topics, refer to the Cisco Press CCNP preparation books found at www.ciscopress.com/ccnp. Cisco LAN Switching, by Kennedy Clark and Kevin Hamilton, covers STP logic and operations in detail. MSTP, PVST+, and Rapid PVST+ (RPVST+) configuration are covered in the “Configuring STP” document at http://www.cisco.com/en/US/docs/switches/lan/catalyst3560/software/release/ 12.2_44_se/configuration/guide/swstp.html.

Blueprint topics covered in this chapter: This chapter covers the following subtopics from the Cisco CCIE Routing and Switching written exam blueprint. Refer to the full blueprint in Table I-1 in the Introduction for more details on the topics covered in each chapter and their context within the blueprint. ■

IPv4 Addressing

■

IPv4 Subnetting

■

IPv4 VLSM

■

Route Summarization

■

NAT

CHAPTER

4

IP Addressing Complete mastery of IP addressing and subnetting is required for any candidate to have a reasonable chance at passing both the CCIE written and lab exam. In fact, even the CCNA exam has fairly rigorous coverage of IP addressing and the related protocols. For the CCIE exam, understanding these topics is required to answer much deeper questions—for instance, a question might ask for the interpretation of the output of a show ip bgp command and a configuration snippet to decide what routes would be summarized into a new prefix. To answer such questions, the basic concepts and math behind subnetting need to be very familiar.

“Do I Know This Already?” Quiz Table 4-1 outlines the major headings in this chapter and the corresponding “Do I Know This Already?” quiz questions. Table 4-1

“Do I Know This Already?” Foundation Topics Section-to-Question Mapping Foundation Topics Section

Questions Covered in This Section

IP Addressing and Subnetting

1–4

CIDR, Private Addresses, and NAT

5–8

Score

Total Score

In order to best use this pre-chapter assessment, remember to score yourself strictly. You can find the answers in Appendix A, “Answers to the ‘Do I Know This Already?’ Quizzes.” 1.

In what subnet does address 192.168.23.197/27 reside? a.

192.168.23.0

b.

192.168.23.128

c.

192.168.23.160

d.

192.168.23.192

e.

192.168.23.196

106

Chapter 4: IP Addressing

2.

3.

4.

5.

Router1 has four LAN interfaces, with IP addresses 10.1.1.1/24, 10.1.2.1/24, 10.1.3.1/24, and 10.1.4.1/24. What is the smallest summary route that could be advertised out a WAN link connecting Router1 to the rest of the network, if subnets not listed here were allowed to be included in the summary? a.

10.1.2.0/22

b.

10.1.0.0/22

c.

10.1.0.0/21

d.

10.1.0.0/16

Router1 has four LAN interfaces, with IP addresses 10.22.14.1/23, 10.22.18.1/23, 10.22.12.1/23, and 10.22.16.1/23. Which one of the answers lists the smallest summary route(s) that could be advertised by R1 without also including subnets not listed in this question? a.

10.22.12.0/21

b.

10.22.8.0/21

c.

10.22.8.0/21 and 10.22.16.0/21

d.

10.22.12.0/22 and 10.22.16.0/22

Which two of the following VLSM subnets, when taken as a pair, overlap? a.

10.22.21.128/26

b.

10.22.22.128/26

c.

10.22.22.0/27

d.

10.22.20.0/23

e.

10.22.16.0/22

Which of the following protocols or tools includes a feature like route summarization, plus administrative rules for global address assignment, with a goal of reducing the size of Internet routing tables? a.

Classless interdomain routing

b.

Route summarization

c.

Supernetting

d.

Private IP addressing

“Do I Know This Already?” Quiz

6.

7.

8.

107

Which of the following terms refers to a NAT feature that allows for significantly fewer IP addresses in the enterprise network as compared with the required public registered IP addresses? a.

Static NAT

b.

Dynamic NAT

c.

Dynamic NAT with overloading

d.

PAT

e.

VAT

Consider an enterprise network using private class A network 10.0.0.0, and using NAT to translate to IP addresses in registered class C network 205.1.1.0. Host 10.1.1.1 has an open www session to Internet web server 198.133.219.25. Which of the following terms refers to the destination address of a packet, sent by the web server back to the client, when the packet has not yet made it back to the enterprise’s NAT router? a.

Inside Local

b.

Inside Global

c.

Outside Local

d.

Outside Global

Router1 has its fa0/0 interface, address 10.1.2.3/24, connected to an enterprise network. Router1’s S0/1 interface connects to an ISP, with the interface using a publicly-registered IP address of 171.1.1.1/30,. Which of the following commands could be part of a valid NAT overload configuration, with 171.1.1.1 used as the public IP address? a.

ip nat inside source list 1 int s0/1 overload

b.

ip nat inside source list 1 pool fred overload

c.

ip nat inside source list 1 171.1.1.1 overload

d.

None of the answers is correct.

108

Chapter 4: IP Addressing

Foundation Topics

IP Addressing and Subnetting You need a postal address to receive letters; similarly, computers must use an IP address to be able to send and receive data using the TCP/IP protocols. Just as the postal service dictates the format and meaning of a postal address to aid the efficient delivery of mail, the TCP/IP protocol suite imposes some rules about IP address assignment so that routers can efficiently forward packets between IP hosts. This chapter begins with coverage of the format and meaning of IP addresses, with required consideration for how they are grouped to aid the routing process.

IP Addressing and Subnetting Review First, here’s a quick review of some of the core facts about IPv4 addresses that should be fairly familiar to you: ■

32-bit binary number.

■

Written in “dotted decimal” notation (for example, 1.2.3.4), with each decimal octet representing 8 bits.

■

Addresses are assigned to network interfaces, so computers or routers with multiple interfaces have multiple IP addresses.

■

A computer with an IP address assigned to an interface is an IP host.

■

A group of IP hosts that are not separated from each other by an IP router are in the same grouping.

■

These groupings are called networks, subnets, or prefixes, depending on the context.

■

IP hosts separated from another set of IP hosts by a router must be in separate groupings (network/subnet/prefix).

IP addresses may be analyzed using classful or classless logic, depending on the situation. Classful logic simply means that the main class A, B, and C rules from RFC 791 are considered. The next several pages present a classful view of IP addresses, as reviewed in Table 4-2. With classful addressing, class A, B, and C networks can be identified as such by their first several bits (shown in the last column of Table 4-1) or by the range of decimal values for their first octets. Also, each class A, B, or C address has two parts (when not subnetted): a network part and a host part. The size of each is implied by the class, and can be stated explicitly using the default mask

IP Addressing and Subnetting

109

for that class of network. For instance, mask 255.0.0.0, the default mask for class A networks, has 8 binary 1s and 24 binary 0s, representing the size of the network and host parts, respectively. Table 4-2

Classful Network Review Class of Address

Size of Network and Host Parts of the Addresses

Range of First Octet Values

Default Mask for Each Class of Network

Identifying Bits at Beginning of Address

A

8/24

1–126

255.0.0.0

0

B

16/16

128–191

255.255.0.0

10

C

24/8

192–223

255.255.255.0

110

D

—

224–239

—

1110

E

—

240–255

—

1111

Subnetting a Classful Network Number With classful addressing, and no subnetting, an entire class A, B, or C network is needed on each individual instance of a data link. For example, Figure 4-1 shows a sample internetwork, with dashed-line circles representing the set of hosts that must be in the same IP network—in this case requiring three networks. Figure 4-1 shows two options for how IP addresses may be assigned and grouped together for this internetwork topology. Figure 4-1

Sample Internetwork with Two Alternatives for Address Assignment—Without and With Subnetting Option 1: Use Classful Networks for Each Group

Network 172.31.0.0

Network 8.0.0.0

Network 130.2.0.0

Client3

E0/0

R3

R1

SW1

172.31.103.41

Subnet 172.31.103.0 255.255.255.0

Subnet 172.31.13.0 255.255.255.0

Option 2: Use Subnets of One Classful Network

Subnet 172.31.11.0 255.255.255.0

110

Chapter 4: IP Addressing

Option 1 uses three classful networks; however, it wastes a lot of IP addresses. For example, all hosts in class A network 8.0.0.0 must reside on the LAN on the right side of the figure. Of course, the much more reasonable alternative is to reserve one classful IP network number, and use subnetting to subdivide that network into at least three subdivisions, called subnets. Option 2 (bottom of Figure 4-1) shows how to subdivide a class A, B, or C network into subnets. To create subnets, the IP addresses must have three fields instead of just two—the network, subnet, and host. When using classful logic to interpret IP addresses, the size of the network part is still defined by classful rules—either 8, 16, or 24 bits based on class. To create the subnet field, the host field is shortened, as shown in Figure 4-2. Figure 4-2

Formats of IP Addresses when Subnetting 8

24 – x

x

Network

Subnet

Host

Class A

16

16 – x

x

Network

Subnet

Host

24 Network

8–x

Class B

x

Subnet Host Class C

NOTE The term internetwork refers to a collection of computers and networking hardware; because TCP/IP discussions frequently use the term network to refer to a classful class A, B, or C IP network, this book uses the term internetwork to refer to an entire network topology, as shown in Figure 4-1.

To determine the size of each field in a subnetted IP address, you can follow the three easy steps shown in Table 4-3. Note that Figure 4-1 also showed alternative addressing for using subnets, with the last column in Table 4-3 showing the size of each field for that particular example, which used class B network 172.31.0.0, mask 255.255.255.0. Table 4-3

Finding the Size of the Network, Subnet, and Host Fields in an IP Address Name of Part of the Address

Process to Find Its Size

Size per Figure 4-1 Example

Network

8, 16, or 24 bits based on class rules

16

Subnet

32 minus network and host bits

8

Host

Equal to the number of binary 0s in the mask

8

IP Addressing and Subnetting

111

Comments on Classless Addressing The terms classless and classful can be applied to three popular topics that are all related to IP. This chapter explains classful and classless IP addressing, which are relatively simple concepts. Two other chapters explain the other uses of the terms classless and classful: Chapter 6, “IP Forwarding (Routing),” describes classless/classful routing, and Appendix E, “RIP Version 2,” covers classless/classful routing protocols. Classless IP addressing, simply put, means that class A, B, and C rules are ignored. Each address is viewed as a two-part address, formally called the prefix and the host parts of the address. The prefix simply states how many of the beginning bits of an IP address identify or define the group. It is the same idea as using the combined network and subnet parts of an address to identify a subnet. All the hosts with identical prefixes are in effect in the same group, which can be called a subnet or a prefix. Just as a classful subnet must be listed with the subnet mask to know exactly which addresses are in the subnet, a prefix must be listed with its prefix length. The prefix itself is a dotteddecimal number. It is typically followed by a / symbol, after which the prefix length is listed. The prefix length is a decimal number that denotes the length (in bits) of the prefix. For example, 172.31.13.0/24 means a prefix of 172.31.13.0 and a prefix length of 24 bits. Also, the prefix can be implied by a subnet mask, with the number of 1s in the binary version of the mask implying the prefix length. Classless and classful addressing are mainly just two ways to think about IP address formats. For the exam, make sure to understand both perspectives and the terminology used by each.

Subnetting Math Knowing how to interpret the meaning of addresses and masks, routes and masks in the routing table, addresses and masks in ACLs, and configure route-filtering are all very important topics for the CCIE Routing and Switching written and lab exams. This section covers the binary math briefly, with coverage of some tricks to do the math quickly without binary math. Several subsequent chapters cover the configuration details of features that require this math.

Dissecting the Component Parts of an IP Address First, deducing the size of the three parts (classful view) or two parts (classless view) of an IP address is important, because it allows you to analyze information about that subnet and other subnets. Every internetwork requires some number of subnets, and some number of hosts per subnet. Analyzing the format of an existing address, based on the mask or prefix length, allows

112

Chapter 4: IP Addressing

you to determine whether enough hosts per subnet exist, or whether enough subnets exist to support the number of hosts. The following list summarizes some of the common math facts about subnetting related to the format of IP addresses: ■

If a subnet has been defined with y host bits, there are 2y – 2 valid usable IP addresses in the subnet, because two numeric values are reserved.

■

One reserved IP address in each subnet is the subnet number itself. This number, by definition, has binary 0s for all host bits. This number represents the subnet, and is typically seen in routing tables.

■

The other reserved IP address in the subnet is the subnet broadcast address, which by definition has binary 1s for all host bits. This number can be used as a destination IP address to send a packet to all hosts in the subnet.

■

When you are thinking classfully, if the mask implies x subnet bits, then 2x possible subnets exist for that classful network, assuming the same mask is used throughout the network.

■

Although there are no truly reserved values for the subnet numbers, two (lowest and highest values) may be discouraged from use in some cases:

— Zero subnet—The subnet field is all binary 0s; in decimal, each zero subnet is the exact same dotted-decimal number as the classful network number, potentially causing confusion.

— Broadcast subnet—The subnet field is all binary 1s; in decimal, this subnet’s broadcast address is the same as the network-wide broadcast address, potentially causing confusion. In Cisco routers, by default, zero subnets and broadcast subnets work fine. You can disable the use of the zero subnet with the no ip subnet-zero global command. The only time that using the zero subnet typically causes problems is when classful routing protocols are used. Finding Subnet Numbers and Valid Range of IP Addresses—Binary When examining an IP address and mask, the process of finding the subnet number, the broadcast address, and the range of valid IP addresses is as fundamental to networking as is addition and subtraction for advanced math. Possibly more so for the CCIE Routing and Switching lab exam, mastery of the math behind subnetting, which is the same basic math behind route summarization and filtering, will improve your speed completing complex configurations on the exam. The range of valid IP addresses in a subnet begins with the number that is one larger than the subnet number, and ends with the address that is one smaller than the broadcast address for the

IP Addressing and Subnetting

113

subnet. So, to determine the range of valid addresses, just calculate the subnet number and broadcast address, which can be done as follows: ■

To derive the subnet number—Perform a bit-wise Boolean AND between the IP address and mask

■

To derive the broadcast address—Change all host bits in the subnet number from 0s to 1s

A bitwise Boolean AND means that you place two long binary numbers on top of each other, and then AND the two bits that line up vertically. (A Boolean AND results in a binary 1 only if both bits are 1; otherwise, the result is 0.) Table 4-4 shows an easy example based on subnet 172.31.103.0/24 from Figure 4-1. Table 4-4

Binary Math to Calculate the Subnet Number and Broadcast Address Address

172.31.103.41

1010 1100 0001 1111 0110 0111 0010 1001

Mask

255.255.255.0

1111 1111 1111 1111 1111 1111 0000 0000

Subnet Number (Result of AND)

172.31.103.0

1010 1100 0001 1111 0110 0111 0000 0000

Broadcast

172.31.103.255

1010 1100 0001 1111 0110 0111 1111 1111

Probably most everyone reading this already knew that the decimal subnet number and broadcast addresses shown in Table 4-4 were correct, even without looking at the binary math. The important part is to recall the binary process, and practice until you can confidently and consistently find the answer without using any binary math at all. The only parts of the math that typically trip people up are the binary to decimal and decimal to binary conversions. When working in binary, keep in mind that you will not have a calculator for the written exam, and that when converting to decimal, you always convert 8 bits at a time—even if an octet contains some prefix bits and some host bits. (Appendix B, “Decimal to Binary Conversion Table,” contains a conversion table for your reference.)

Decimal Shortcuts to Find the Subnet Number and Valid Range of IP Addresses Many of the IP addressing and routing related problems on the exam come back to the ability to solve a couple of classic basic problems. One of those problems runs as follows: Given an IP address and mask (or prefix length), determine the subnet number/prefix, broadcast address, and range of valid IP addresses. If you personally can already solve such problems with only a few seconds’ thought, even with tricky masks, then you can skip this section of the chapter. If you cannot solve such questions

114

Chapter 4: IP Addressing

easily and quickly, this section can help you learn some math shortcuts that allow you to find the answers without needing to use any Boolean math. NOTE The next several pages of this chapter describe some algorithms you can use to find many important details related to IP addressing, without needing to convert to and from binary. In my experience, some people simply work better performing the math in binary until the answers simply start popping into their heads. Others find that the decimal shortcuts are more effective. If you use the decimal shortcuts, it is best to practice them until you no longer really use the exact steps listed in this book; rather, the processes should become second nature. To that end, CD-only Appendix D, “IP Addressing Practice,” lists several practice problems for each of the algorithms presented in this chapter.

To solve the “find the subnet/broadcast/range of addresses” type of problem, at least three of the four octets should have pretty simple math. For example, with a nice, easy mask like 255.255.255.0, the logic used to find the subnet number and broadcast address is intuitive to most people. The more challenging cases occur when the mask or prefix does not divide the host field at a byte boundary. For instance, the same IP address 172.31.103.41, with mask 255.255.252.0 (prefix /22), is actually in subnet 172.31.100.0. Working with the third octet in this example is the hard part, because the mask value for that octet is not 0 or 255; for the upcoming process, this octet is called the interesting octet. The following process finds the subnet number, using decimal math, even with a challenging mask: Step 1

Find the mask octets of value 255; copy down the same octets from the IP address.

Step 2

Find the mask octets of value 0; write down 0s for the same octets.

Step 3

If one octet has not yet been filled in, that octet is the interesting octet. Find the subnet mask’s value in the interesting octet, and subtract it from 256. Call this number the “magic number.”

Step 4

Find the integer multiple of the magic number that is closest to, but not larger than, the interesting octet’s value.

An example certainly helps, as shown in Table 4-5, with 172.31.103.41, mask 255.255.252.0. The table separates the address into its four component octets. In this example, the first, second, and fourth octets of the subnet number are easily found from Steps 1 and 2 in the process. Because the interesting octet is the third octet, the magic number is 256 – 252, or 4. The integer multiple of 4, closest to 103 but not exceeding 103, is 100—making 100 the subnet number’s value in the third octet. (Note that you can use this same process even with an easy mask, and Steps 1 and 2 will give you the complete subnet number.)

IP Addressing and Subnetting

Table 4-5

115

Quick Math to Find the Subnet Number—172.31.103.41, 255.255.252.0 Octet

Comments

1

2

3

4

Address

172

31

103

41

Mask

255

255

252

0

Equivalent to /22.

Subnet number results after Steps 1 and 2

172

31

0

Magic number will be 256 – 252 = 4.

Subnet number after completing the interesting octet

172

31

0

100 is the multiple of 4 closest to, but not exceeding, 103.

100

A similar process can be used to determine the subnet broadcast address. This process assumes that the mask is tricky. The detailed steps are as follows: Step 1

Start with the subnet number.

Step 2

Decide which octet is interesting, based on which octet of the mask does not have a 0 or 255.

Step 3

For octets to the left of the interesting octet, copy down the subnet number’s values into the place where you are writing down the broadcast address.

Step 4

For any octets to the right of the interesting octet, write down 255 for the broadcast address.

Step 5

Calculate the magic number: find the subnet mask’s value in the interesting octet and subtract it from 256.

Step 6

Take the subnet number’s interesting octet value, add the magic number to it, and subtract 1. Fill in the broadcast address’s interesting octet with this number.

Table 4-6 shows the 172.31.103.41/22 example again, using this process to find the subnet broadcast address. Table 4-6

Quick Math to Find the Broadcast Address—172.31.103.41, 255.255.252.0 Octet

Comments

1

2

3

4

Subnet number (per Step 1)

172

31

100

0

Mask (for reference)

255

255

252

0

Results after Steps 1 to 4

172

31

Subnet number after completing the empty octet

172

31

103

Equivalent to /22

255

Magic number will be 256 – 252 = 4

255

Subnet’s third octet (100), plus magic number (4), minus 1 is 103

116

Chapter 4: IP Addressing

NOTE If you have read the last few pages to improve your speed at dissecting a subnet without requiring binary math, it is probably a good time to pull out the CD in the back of the book. CDonly Appendix D, “IP Addressing Practice,” contains several practice problems for finding the subnet and broadcast address, as well as for many other math related to IP addressing.

Determining All Subnets of a Network—Binary Another common question, typically simply a portion of a more challenging question on the CCIE written exam, relates to finding all subnets of a network. The base underlying question might be as follows: Given a particular class A, B, or C network, and a mask/prefix length used on all subnets of that network, what are the actual subnet numbers? The answers can be found using binary or using a simple decimal algorithm. This section first shows how to answer the question using binary, using the following steps. Note that the steps include details that are not really necessary for the math part of the problem; these steps are mainly helpful for practicing the process. Step 1

Write down the binary version of the classful network number; that value is actually the zero subnet as well.

Step 2

Draw two vertical lines through the number, one separating the network and subnet parts of the number, the other separating the subnet and host part.

Step 3

Calculate the number of subnets, including the zero and broadcast subnet, based on 2y, where y is the number of subnet bits.

Step 4

Write down y-1 copies of the binary network number below the first one, but leave the subnet field blank.

Step 5

Using the subnet field as a binary counter, write down values, top to bottom, in which the next value is 1 greater than the previous.

Step 6

Convert the binary numbers, 8 bits at a time, back to decimal.

This process takes advantage of a couple of facts about the binary form of IP subnet numbers: ■

All subnets of a classful network have the same value in the network portion of the subnet number.

■

All subnets of any classful network have binary 0s in the host portion of the subnet number.

Step 4 in the process simply makes you write down the network and host parts of each subnet number, because those values are easily predicted. To find the different subnet numbers, you then

IP Addressing and Subnetting

117

just need to discover all possible different combinations of binary digits in the subnet field, because that is the only part of the subnet numbers that differs from subnet to subnet. For example, consider the same class B network 172.31.0.0, with static length subnet masking (SLSM) assumed, and a mask of 255.255.224.0. Note that this example uses 3 subnet bits, so there will be 23 subnets. Table 4-7 lists the example. Table 4-7

Binary Method to Find All Subnets—Steps 1 Through 4 Octet Subnet

1

2

3

Network number/zero subnet

10101100

000 11111

2nd subnet

10101100

3rd subnet

000

4

00000

00000000

000 11111

00000

00000000

10101100

000 11111

00000

00000000

4th subnet

10101100

000 11111

00000

00000000

5th subnet

10101100

000 11111

00000

00000000

6th subnet

10101100

000 11111

00000

00000000

7th subnet

10101100

000 11111

00000

00000000

8th subnet (2y = 8); broadcast subnet

10101100

000 11111

00000

00000000

At this point, you have the zero subnet recorded at the top, and you are ready to use the subnet field (the missing bits in the table) as a counter to find all possible values. Table 4-8 completes the process. Table 4-8

Binary Method to Find All Subnets—Step 5 Octet Subnet

1

2

3

4

Network number/zero subnet

10101100

00011111

000

00000

00000000

2nd subnet

10101100

00011111

001

00000

00000000

3rd subnet

10101100

00011111

010

00000

00000000

4th subnet

10101100

00011111

011

00000

00000000

5th subnet

10101100

00011111

100

00000

00000000

6th subnet

10101100

00011111

101

00000

00000000

7th subnet

10101100

00011111

110

00000

00000000

8th subnet (2y = 8); broadcast subnet

10101100

00011111

111

00000

00000000

118

Chapter 4: IP Addressing

The final step to determine all subnets is simply to convert the values back to decimal. Take care to always convert 8 bits at a time. In this case, you end up with the following subnets: 172.31.0.0, 172.31.32.0, 172.31.64.0, 172.31.96.0, 172.31.128.0, 172.31.160.0, 172.31.192.0, and 172.31.224.0. Determining All Subnets of a Network—Decimal You may have noticed the trend in the third octet values in the subnets listed in the previous paragraph. When assuming SLSM, the subnet numbers in decimal do have a regular increment value, which turns out to be the value of the magic number. For example, instead of the binary math in the previous section, you could have thought the following: ■

The interesting octet is the third octet.

■

The magic number is 256 – 224 = 32.

■

172.31.0.0 is the zero subnet, because it is the same number as the network number.

■

The other subnet numbers are increments of the magic number inside the interesting octet.

If that logic already clicks in your head, you can skip to the next section in this chapter. If not, the rest of this section outlines an decimal algorithm that takes a little longer pass at the same general logic. First, the question and the algorithm assume that the same subnet mask is used on all subnets of this one classful network—a feature sometimes called static length subnet masking (SLSM). In contrast, variable length subnet masking (VLSM) means that different masks are used in the same classful network. The algorithm assumes a subnet field of 8 bits or less just to keep the steps uncluttered; for longer subnet fields, the algorithm can be easily extrapolated. Step 1

Write down the classful network number in decimal.

Step 2

For the first (lowest numeric) subnet number, copy the entire network number. That is the first subnet number, and is also the zero subnet.

Step 3

Decide which octet contains the entire subnet field; call this octet the interesting octet. (Remember, this algorithm assumes 8 subnet bits or less, so the entire subnet field will be in a single interesting octet.)

Step 4

Calculate the magic number by subtracting the mask’s interesting octet value from 256.

Step 5

Copy down the previous subnet number’s noninteresting octets onto the next line as the next subnet number; only one octet is missing at this point.

Step 6

Add the magic number to the previous subnet’s interesting octet, and write that down as the next subnet number’s interesting octet, completing the next subnet number.

IP Addressing and Subnetting

Step 7

119

Repeat Steps 5 and 6 until the new interesting octet is 256. That subnet is not valid. The previously calculated subnet is the last valid subnet, and also the broadcast subnet.

For example, consider the same class B network 172.31.0.0, with SLSM assumed, and a mask of 255.255.224.0. Table 4-9 lists the example. Table 4-9

Subnet List Chart—172.31.0.0/255.255.224.0 Octet

Comments

1

2

3

4

Network number

172

31

0

0

Step 1 from the process.

Mask

255

255

224

0

Magic number is 256 – 224 = 32.

Subnet zero

172

31

0

0

Step 2 from the process.

First subnet

172

31

32

0

Steps 5 and 6; previous interesting octet 0, plus magic number (32).

Next subnet

172

31

64

0

32 plus magic number is 64.

Next subnet

172

31

96

0

64 plus magic number is 96.

Next subnet

172

31

128

0

96 plus magic number is 128.

Next subnet

172

31

160

0

128 plus magic number is 160.

Next subnet

172

31

192

0

160 plus magic number is 192.

Last subnet (broadcast)

172

31

224

0

The broadcast subnet in this case.

Invalid; easy-to-recognize stopping point

172

31

256

0

256 is out of range; when writing this one down, note that it is invalid, and that the previous one is the last valid subnet.

You can use this process repeatedly as needed until the answers start jumping out at you without the table and step-wise algorithm. For more practice, refer to CD-only Appendix D.

VLSM Subnet Allocation So far in this chapter, most of the discussion has been about examining existing addresses and subnets. Before deploying new networks, or new parts of a network, you must give some thought to the ranges of IP addresses to be allocated. Also, when assigning subnets for different locations, you should assign the subnets with thought for how routes could then be summarized. This section covers some of the key concepts related to subnet allocation and summarization. (This section focuses on the concepts behind summarization; the configuration of route summarization is routing protocol–specific and thus is covered in the individual chapters covering routing protocols.)

120

Chapter 4: IP Addressing

Many organizations purposefully use SLSM to simplify operations. Additionally, many internetworks also use private IP network 10.0.0.0, with an SLSM prefix length of /24, and use NAT for connecting to the Internet. By using SLSM, particularly with a nice, easy prefix like /24, operations and troubleshooting can be a lot easier. In some cases, VLSM is required or preferred when allocating addresses. VLSM is typically chosen when the address space is constrained to some degree. The VLSM subnet assignment strategy covered here complies with the strategy you may remember from the Cisco BSCI course or from reading the Cisco Press CCNP Routing certification books. Similar to when assigning subnets with SLSM, you should use an easily summarized block of addresses for a new part of the network. Because VLSM network addresses are likely constrained to some degree, you should choose the specific subnets wisely. The general rules for choosing wisely are as follows: Step 1

Determine the shortest prefix length (in other words, the largest block) required.

Step 2

Divide the available address block into equal-sized prefixes based on the shortest prefix from Step 1.

Step 3

Allocate the largest required subnets/prefixes from the beginning of the IP address block, leaving some equal-sized unallocated address blocks at the end of the original large address block.

Step 4

Choose an unallocated block that you will further subdivide by repeating the first three steps, using the shortest required prefix length (largest address block) for the remaining subnets.

Step 5

When allocating very small address blocks for use on links between routers, consider using subnets at the end of the address range. This leaves the largest consecutive blocks available in case future requirements change.

For instance, imagine that a network engineer plans a new site installation. He allocates the 172.31.28.0/23 address block for the new site, expecting to use the block as a single summarized route. When planning, the engineer then subdivides 172.31.28.0/23 per the subnet requirements for the new installation, as shown in Figure 4-3. The figure shows three iterations through the VLSM subnet assignment process, because the requirements call for three different subnet sizes. Each iteration divides a remaining block into equal sizes, based on the prefix requirements of the subnets allocated at that step. Note that the small /30 prefixes were allocated from the end of the address range, leaving the largest possible consecutive address range for future growth.

IP Addressing and Subnetting

Figure 4-3

121

Example of VLSM Subnet Allocation Process 172.31.28.0.0/23 (172.31.28.0 Through 172.31.29.255) Requirements: 3 /25’s 2 /27’s 3 /30’s Pass 1: /25 prefixes Block 172.31.28.0/23

Allocated

Allocated

Allocated

Unallocated

172.31.28.0/25

172.31.28.128/25

172.31.29.0/25

172.31.29.128/25

172.31.28.1 – 172.31.28.126

172.31.28.129 – 172.31.28.254

172.31.29.1 – 172.31.29.126

172.31.29.129 – 172.31.29.254

Pass 2: /27 prefixes Block 172.31.29.128/25

Allocated

Allocated

172.31.29.128/27

172.31.29.160/27

Step 3: /30 prefixes Allocate High End:

Unallocated

Unallocated Unallocated 172.31.29.192/27

172.31.29.224/27

Allocated

172.31.29.252/30, 172.31.29.248/30, 172.31.29.244/30

Route Summarization Concepts The ability to recognize and define how to most efficiently summarize existing address ranges is an important skill on both the written and lab exams. For the written exam, the question may not be as straightforward as, “What is the most efficient summarization of the following subnets?” Rather, the math required for such a question might simply be part of a larger question. Certainly, such math is required for the lab exam. This section looks at the math behind finding the best summarization; other chapters cover specific configuration commands. Good IP address assignment practices should always consider the capabilities for route summarization. For instance, if a division of a company needs 15 subnets, an engineer needs to allocate those 15 subnets from the unused portions of the address block available to that internetwork. However, assigning subnets 10.1.101.0/24 through 10.1.115.0/24 would be a poor choice, because those do not easily summarize. Rather, allocate a range of addresses that can be easily summarized into a single route. For instance, subnets 10.1.96.0/24 through 10.1.110.0/24 can be summarized as a single 10.1.96.0/20 route, making those routes a better choice. There are two main ways to think of the word “best” when you are looking for the “best summarization”: ■

Inclusive summary routes—A single summarized route that is as small a range of addresses as possible, while including all routes/subnets shown, and possibly including subnets that do not currently exist.

122

Chapter 4: IP Addressing

■

Exclusive summary routes—As few as possible summarized routes that include all to-besummarized address ranges, but excluding all other routes/subnets.

NOTE The terms inclusive summary, exclusive summary, and candidate summary are simply terms I invented for this book and will continue to use later in the chapter. For instance, with the VLSM example in Figure 4-3, the network engineer purposefully planned so that an inclusive summary of 172.31.28.0/23 could be used. Even though not all subnets are yet allocated from that address range, the engineer is likely saving the rest of that address range for future subnets at that site, so summarizing using an inclusive summary is reasonable. In other cases, typically when trying to summarize routes in an internetwork for which summarization was not planned, the summarization must exclude routes that are not explicitly listed, because those address ranges may actually be used in another part of the internetwork. Finding Inclusive Summary Routes—Binary Finding the best inclusive summary lends itself to a formal binary process, as well as to a formal decimal process. The binary process runs as follows: Step 1

Write down the binary version of each component subnet, one on top of the other.

Step 2

Inspect the binary values to find how many consecutive bits have the exact same value in all component subnets. That number of bits is the prefix length.

Step 3

Write a new 32-bit number at the bottom of the list by copying y bits from the prior number, y being the prefix length. Write binary 0s for the remaining bits. This is the inclusive summary.

Step 4

Convert the new number to decimal, 8 bits at a time.

Table 4-10 shows an example of this process, using four routes, 172.31.20.0, .21.0, .22.0, and .23.0, all with prefix /24. The second example adds 172.31.24.0 to that same list. Table 4-10

Example of Finding the Best Inclusive Summary—Binary Octet 1

Octet 2

Octet 3

Octet 4

172.31.20.0/24

10101100

00011111

000101

00

00000000

172.31.21.0/24

10101100

00011111

000101

01

00000000

172.31.22.0/24

10101100

00011111

000101

10

00000000

172.31.23.0/24

10101100

00011111

000101

11

00000000

10101100

00011111

000101

00

00000000

Prefix length: 22 Inclusive summary

IP Addressing and Subnetting

123

The trickiest part is Step 2, in which you have to simply look at the binary values and find the point at which the bits are no longer equal. You can shorten the process by, in this case, noticing that all component subnets begin with 172.31, meaning that the first 16 bits will certainly have the same values. Finding Inclusive Summary Routes—Decimal To find the same inclusive summary using only decimal math, use the following process. The process works just fine with variable prefix lengths and nonconsecutive subnets. Step 1

Count the number of subnets; then, find the smallest value of y, such that 2y => that number of subnets.

Step 2

For the next step, use a prefix length based on the longest prefix length of the component subnets, minus y.

Step 3

Pretend that the lowest numeric subnet number in the list of component subnets is an IP address. Using the new, smaller prefix from Step 2, calculate the subnet number in which this pretend address resides.

Step 4

Repeat Step 3 for the largest numeric component subnet number and the same prefix. If it is the same subnet derived as in Step 3, the resulting subnet is the best summarized route, using the new prefix.

Step 5

If Steps 3 and 4 do not yield the same resulting subnet, repeat Steps 3 and 4 with another new prefix length of 1 less than the last prefix length.

Table 4-11 shows two examples of the process. The first example has four routes, 172.31.20.0, .21.0, .22.0, and .23.0, all with prefix /24. The second example adds 172.31.24.0 to that same list. Table 4-11

Example of Finding the Best Summarizations Step

Range of .20.0, .21.0, .22.0, and .23.0, /24

Same Range, Plus 172.31.24.0

Step 1

22 = 4, y = 2

23 = 8, y = 3

Step 2

24 – 2 = 22

24 – 3 = 21

Step 3

Smallest subnet 172.31.20.0, with /22, yields 172.31.20.0/22

Smallest subnet 172.31.20.0, with /21, yields 172.31.16.0/21

Step 4

Largest subnet 172.31.23.0, with /22, yields 172.31.20.0/22

Largest subnet 172.31.24.0, with /21, yields 172.31.24.0/21

Step 5

—

21 – 1 = 20; new prefix

Step 3, 2nd time

—

172.31.16.0/20

Step 4, 2nd time

—

172.31.16.0/20; the same as prior step, so that is the answer

124

Chapter 4: IP Addressing

With the first example, Steps 3 and 4 yielded the same answer, which means that the best inclusive summary had been found. With the second example, a second pass through the process was required. CD-only Appendix D contains several practice problems to help you develop speed and make this process second nature. Finding Exclusive Summary Routes—Binary A similar process, listed next, can be used to find the exclusive summary. Keep in mind that the best exclusive summary can be comprised of multiple summary routes. Once again, to keep it simple, the process assumes SLSM. Step 1

Find the best exclusive summary route; call it a candidate exclusive summary route.

Step 2

Determine if the candidate summary includes any address ranges it should not. To do so, compare the summary’s implied address range with the implied address ranges of the component subnets.

Step 3

If the candidate summary only includes addresses in the ranges implied by the component subnets, the candidate summary is part of the best exclusive summarization of the original component subnets.

Step 4

If instead the candidate summary includes some addresses that match the candidate summary routes and some addresses that do not, split the current candidate summary in half, into two new candidate summary routes, each with a prefix 1 longer than before.

Step 5

If the candidate summary only includes addresses outside the ranges implied by the component subnets, the candidate summary is not part of the best exclusive summarization, and it should not be split further.

Step 6

Repeat Steps 2 through 4 for each of the two possible candidate summary routes created at Step 4.

For example, take the same five subnets used with the inclusive example—172.31.20.0/24, .21.0, .22.0, .23.0, and .24.0. The best inclusive summary is 172.31.16.0/20, which implies an address range of 172.31.16.0 to 172.31.31.255—clearly, it includes more addresses than the original five subnets. So, repeat the process of splitting the summary in half, and repeating, until summaries are found that do not include any unnecessary address ranges. Figure 4-4 shows the idea behind the logic. The process starts with one candidate summary. If it includes some addresses that need to be summarized and some addresses it should not summarize, split it in half, and try again with each half. Eventually, the best exclusive summary routes are found, or the splitting keeps happening until you get back to the original routes. In fact, in this case, after a few more splits (not shown), the process ends up splitting to 172.31.24.0/24, which is one of the original routes—meaning that 172.31.24.0/24 cannot be summarized any further in this example.

CIDR, Private Addresses, and NAT

Figure 4-4

125

Example of Process to Find Exclusive Summary Routes

Routes to Summarize: 172.31.20.0/24 (20.0 thru 20.255) 172.31.21.0/24 (21.0 thru 21.255) 172.31.22.0/24 (22.0 thru 22.255) 172.31.23.0/24 (23.0 thru 23.255) 172.31.24.0/24 (24.0 thru 24.255)

172.31.16.0/20: 16.0 Thru 31.255 Too Inclusive: Split!

172.31.16.0/21: 16.0 thru 23.255 Too Inclusive: Split!

172.31.24.0/21: 24.0 thru 31.255 Too Inclusive: Split!

172.31.16.0/22: 16.0 Thru 19.255

172.31.20.0/22: 20.0 Thru 23.255

172.31.24.0/22: 24.0 Thru 27.255

172.31.28.0/22: 28.0 Thru 31.255

Range completely outside range to be summarized; stop splitting.

Range is exclusively from target range – keep this as part of best exclusive summary!

Too inclusive: keep splitting! (Details not shown.)

Range completely outside range to be summarized; stop splitting.

CIDR, Private Addresses, and NAT The sky was falling in the early 1990s in that the commercialization of the Internet was rapidly depleting the IP Version 4 address space. Also, Internet routers’ routing tables were doubling annually (at least). Without some changes, the incredible growth of the Internet in the 1990s would have been stifled. To solve the problems associated with this rapid growth, several short-term solutions were created, as well as an ultimate long-term solution. The short-term solutions included classless interdomain routing (CIDR), which helps reduce the size of routing tables by aggregating routes, and Network Address Translation (NAT), which reduces the number of required public IP addresses used by each organization or company. This section covers the details of CIDR and NAT, plus a few related features. The long-term solution to this problem, IPv6, is covered in Chapter 20, “IP Version 6.”

Classless Interdomain Routing CIDR is a convention defined in RFCs 1517 through 1520 that calls for aggregating routes for multiple classful network numbers into a single routing table entry. The primary goal of CIDR is to improve the scalability of Internet routers’ routing tables. Imagine the implications of an Internet router being burdened by carrying a route to every class A, B, and C network on the planet! CIDR uses both technical tools and administrative strategies to reduce the size of the Internet routing tables. Technically, CIDR uses route summarization, but with Internet scale in mind.

126

Chapter 4: IP Addressing

For instance, CIDR might be used to allow a large ISP to control a range of IP addresses from 198.0.0.0 to 198.255.255.255, with the improvements to routing shown in Figure 4-5. Figure 4-5

Typical Use of CIDR

Route to 198.0.0.0 Mask 255.0.0.0 Points to ISP #1

ISP #2 Customer #1 198.8.3.0/24

Route to 198.0.0.0 Mask 255.0.0.0 Points to ISP #1

ISP #3

ISP #1 198.0.0.0 198.255.255.0 Customer #2 198.4.2.0/24 198.4.3.0/24

Route to 198.0.0.0 Mask 255.0.0.0 Points to ISP #1

ISP #4 Customer #3 198.1.0.0

ISPs 2, 3, and 4 need only one route (198.0.0.0/8) in their routing tables to be able to forward packets to all destinations that begin with 198. Note that this summary actually summarizes multiple class C networks—a typical feature of CIDR. ISP 1’s routers contain more detailed routing entries for addresses beginning with 198, based on where they allocate IP addresses for their customers. ISP 1 would reduce its routing tables similarly with large ranges used by the other ISPs. CIDR attacks the problem of large routing tables via administrative means as well. As shown in Figure 4-5, ISPs are assigned contiguous blocks of addresses to use when assigning addresses for their customers. Likewise, regional authorities are assigned large address blocks, so when individual companies ask for registered public IP addresses, they ask their regional registry to assign them an address block. As a result, addresses assigned by the regional agency will at least be aggregatable into one large geographic region of the world. For instance, the Latin American and Caribbean Internet Addresses Registry (LACNIC, http://www.lacnic.net) administers the IP address space of the Latin American and Caribbean region (LAC) on behalf of the Internet community. In some cases, the term CIDR is used a little more generally than the original intent of the RFCs. Some texts use the term CIDR synonymously with the term route summarization. Others use the term CIDR to refer to the process of summarizing multiple classful networks together. In other cases, when an ISP assigns subsets of a classful network to a customer who does not need an entire class C network, the ISP is essentially performing subnetting; once again, this idea sometimes gets categorized as CIDR. But CIDR itself refers to the administrative assignment of large address blocks, and the related summarized routes, for the purpose of reducing the size of the Internet routing tables.

CIDR, Private Addresses, and NAT

127

NOTE Because CIDR defines how to combine routes for multiple classful networks into a single route, some people think of this process as being the opposite of subnetting. As a result, many people refer to CIDR’s summarization results as supernetting.

Private Addressing One of the issues with Internet growth was the assignment of all possible network numbers to a small number of companies or organizations. Private IP addressing helps to mitigate this problem by allowing computers that will never be directly connected to the Internet to not use public, Internet-routable addresses. For IP hosts that will purposefully have no direct Internet connectivity, you can use several reserved network numbers, as defined in RFC 1918 and listed in Table 4-12. Table 4-12

RFC 1918 Private Address Space Range of IP Addresses

Class of Networks

Number of Networks

10.0.0.0 to 10.255.255.255

A

1

172.16.0.0 to 172.31.255.255

B

16

192.168.0.0 to 192.168.255.255

C

256

In other words, any organization can use these network numbers. However, no organization is allowed to advertise these networks using a routing protocol on the Internet. Furthermore, all Internet routers should be configured to reject these routes.

Network Address Translation NAT, defined in RFC 1631, allows a host that does not have a valid registered IP address to communicate with other hosts on the Internet. NAT has gained such wide-spread acceptance that the majority of enterprise IP networks today use private IP addresses for most hosts on the network and use a small block of public IP addresses, with NAT translating between the two. NAT translates, or changes, one or both IP addresses inside a packet as it passes through a router. (Many firewalls also perform NAT; for the CCIE Routing and Switching exam, you do not need to know NAT implementation details on firewalls.) In most cases, NAT changes the (typically private range) addresses used inside an enterprise network into address from the public IP address space. For instance, Figure 4-6 shows static NAT in operation; the enterprise has registered class C network 200.1.1.0/24, and uses private class A network 10.0.0.0/8 for the hosts inside its network.

128

Chapter 4: IP Addressing

Figure 4-6

Basic NAT Concept SA 200.1.1.1

SA 10.1.1.1

10.1.1.1

Server e0/0 NAT

s0/0

Internet 170.1.1.1

10.1.1.2

Outside

Inside DA 10.1.1.1 Inside Local 10.1.1.1 10.1.1.2

DA 200.1.1.1 Inside Global 200.1.1.1 200.1.1.2

Beginning with the packets sent from a PC on the left to the server on the right, the private IP source address 10.1.1.1 is translated to a public IP address of 200.1.1.1. The client sends a packet with source address 10.1.1.1, but the NAT router changes the source to 200.1.1.1—a registered public IP address. Once the server receives a packet with source IP address 200.1.1.1, the server thinks it is talking to host 200.1.1.1, so it replies with a packet sent to destination 200.1.1.1. The NAT router then translates the destination address (200.1.1.1) back to 10.1.1.1. Figure 4-6 provides a good backdrop for the introduction of a couple of key terms, Inside Local and Inside Global. Both terms take the perspective of the owner of the enterprise network. In Figure 4-6, address 10.1.1.1 is the Inside Local address, and 200.1.1.1 is the Inside Global address. Both addresses represent the client PC on the left, which is inside the enterprise network. Address 10.1.1.1 is from the enterprise’s IP address space, which is only locally routable inside the enterprise—hence the term Inside Local. Address 200.1.1.1 represents the local host, but the address is from the globally routable public IP address space—hence the name Inside Global. Table 4-13 lists and describes the four main NAT address terms. Static NAT Static NAT works just like the example in Figure 4-6, but with the IP addresses statically mapped to each other via configuration commands. With static NAT: ■

A particular Inside Local address always maps to the same Inside Global (public) IP address.

CIDR, Private Addresses, and NAT

129

NAT Terminology

Table 4-13

Location of Host Represented by Address

IP Address Space in Which Address Exists

Inside Local address

Inside the enterprise network

Part of the enterprise IP address space; typically a private IP address

Inside Global address

Inside the enterprise network

Part of the public IP address space

Outside Local address

In the public Internet; or, outside the enterprise network

Part of the enterprise IP address space; typically a private IP address

Outside Global address

In the public Internet; or, outside the enterprise network

Part of the public IP address space

Name

■

If used, each Outside Local address always maps to the same Outside Global (public) IP address.

■

Static NAT does not conserve public IP addresses.

Although static NAT does not help with IP address conservation, static NAT does allow an engineer to make an inside server host available to clients on the Internet, because the inside server will always use the same public IP address. Example 4-1 shows a basic static NAT configuration based on Figure 4-6. Conceptually, the NAT router has to identify which interfaces are inside (attach to the enterprise’s IP address space) or outside (attach to the public IP address space). Also, the mapping between each Inside Local and Inside Global IP address must be made. (Although not needed for this example, outside addresses can also be statically mapped.)

Example 4-1

Static NAT Configuration

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! ! E0/0 attaches to the internal Private IP space, so it is configured as an inside ! interface. interfac e Ethernet0/0 ip address 10.1.1.3 255.255.255.0 ip na t inside ! S0/0 is attached to the public Internet, so it is defined as an outside ! interface. interface Seria l0/0 ip address 200.1.1.251 255.255.255.0 i p nat outside

continues

130

Chapter 4: IP Addressing

Example 4-1

Static NAT Configuration (Continued)

! Next, two inside addresses are mapped, with the first address stating the ! Inside Local address, and the next stating the Inside Global address. i p n a t i n s i d e s o u r c e s t a ti c 1 0 . 1 . 1 . 2 2 0 0 . 1 . 1 . 2 ip nat inside source static 10.1.1.1 200.1.1.1 ! Below, the NAT table lists the permanent static entries from the configuration. NAT# show ip nat translations Pro Inside global

Inside local

Outside local

Outside global

--- 200.1.1.1

10.1.1.1

---

--

--- 200.1.1.2

10.1.1.2

---

---

The router is performing NAT only for inside addresses. As a result, the router processes packets entering E0/0—packets that could be sent by inside hosts—by examining the source IP address. Any packets with a source IP address listed in the Inside Local column of the show ip nat translations command output (10.1.1.1 or 10.1.1.2) will be translated to source address 200.1.1.1 or 200.1.1.2, respectively, per the NAT table. Likewise, the router examines the destination IP address of packets entering S0/0, because those packets would be destined for inside hosts. Any such packets with a destination of 200.1.1.1 or .2 will be translated to 10.1.1.1 or .2, respectively. In cases with static outside addresses being configured, the router also looks at the destination IP address of packets sent from the inside to the outside interfaces, and the source IP address of packets sent from outside interfaces to inside interfaces. Dynamic NAT Without PAT Dynamic NAT (without PAT), like static NAT, creates a one-to-one mapping between an Inside Local and Inside Global address. However, unlike static NAT, it does so by defining a set or pool of Inside Local and Inside Global addresses, and dynamically mapping pairs of addresses as needed. For example, Figure 4-7 shows a pool of five Inside Global IP addresses—200.1.1.1 through 200.1.1.5. NAT has also been configured to translate any Inside Local addresses whose address starts with 10.1.1. The numbers 1, 2, and 3 in Figure 4-7 refer to the following sequence of events: 1.

Host 10.1.1.2 starts by sending its first packet to the server at 170.1.1.1.

2.

As the packet enters the NAT router, the router applies some matching logic to decide if the packet should have NAT applied. Because the logic has been configured to mean “translate Inside Local addresses that start with 10.1.1,” the router dynamically adds an entry in the NAT table for 10.1.1.2 as an Inside Local address.

CIDR, Private Addresses, and NAT

Figure 4-7

131

Dynamic NAT 1

4 SA 200.1.1.1

SA 10.1.1.2

Outside

Inside

Server

10.1.1.1 Internet NAT

170.1.1.1

10.1.1.2 NAT Table Before First Packet Criteria for Hosts to NAT: 10.1.1.0 - 10.1.1.255 2

Inside Local

NAT Table After First Packet Inside Local 10.1.1.2

3.

Inside Global

Inside Global 200.1.1.1

3

NAT Pool: 200.1.1.1 200.1.1.2 200.1.1.3 200.1.1.4 200.1.1.5

The NAT router needs to allocate a corresponding IP address from the pool of valid Inside Global addresses. It picks the first one available (200.1.1.1 in this case) and adds it to the NAT table to complete the entry.

With the completion of step 3, the NAT router can actually translate the source IP address, and forward the packet. Note that as long as the dynamic NAT entry exists in the NAT table, only host 10.1.1.2 can use Inside Global IP address 200.1.1.1.

Overloading NAT with Port Address Translation As mentioned earlier, NAT is one of the key features that helped to reduce the speed at which the IPv4 address space was being depleted. NAT overloading, also known as Port Address Translation (PAT), is the NAT feature that actually provides the significant savings of IP addresses. The key to understanding how PAT works is to consider the following: From a server’s perspective, there is no significant difference between 100 different TCP connections, each from a different host, and 100 different TCP connections all from the same host. PAT works by making large numbers of TCP or UDP flows from many Inside Local hosts appear to be the same number of large flows from one (or a few) host’s Inside Global addresses. With PAT,

132

Chapter 4: IP Addressing

instead of just translating the IP address, NAT also translates the port numbers as necessary. And because the port number fields are 16 bits in length, each Inside Global IP address can support over 65,000 concurrent TCP and UDP flows. For instance, in a network with 1000 hosts, a single public IP address used as the only Inside Global address could handle an average of six concurrent flows from each host to and from hosts on the Internet. Dynamic NAT and PAT Configuration Like static NAT, dynamic NAT configuration begins with identifying the inside and outside interfaces. Additionally, the set of Inside Local addresses is configured with the ip nat inside global command. If you are using a pool of public Inside Global addresses, the set of addresses is defined by the ip nat pool command. Example 4-2 shows a dynamic NAT configuration based on the internetwork shown in Figure 4-7. The example defines 256 Inside Local addresses and two Inside Global addresses. Example 4-2

Dynamic NAT Configuration

! First, the ip nat pool fred command lists a range of IP addresses. The ip nat ! inside source list 1 pool fred command points to ACL 1 as the list of Inside ! Local addresses, with a cross-reference to the pool name. interface Ethernet0/0 ip address 10 .1.1.3 255.255.255.0 ip na t inside ! interface Seri al0/0 ip address 200.1.1.251 255.255.255.0 ip nat outside ! ip nat pool fred 200.1.1.1 200.1.1.2 netmask 255.255.255.252 i p n a t i n s i d e so u r c e l i s t 1 p o o l f r e d ! access-list 1 permit 10.1.1.0 0.0.0.255 ! Next, the NAT table begins as an empty table, because no dynamic entries had ! been created at that point. NAT# show ip nat translations ! The NAT statistics show that no hits or misses have occurred. Hits occur when ! NAT looks for a mapping, and finds one. Misses occur when NAT looks for a NAT ! table entry, does not find one, and then needs to dynamically add one. NAT# show ip nat statistics Total active translations: 0 (0 static, 0 dynamic; 0 extended) Outside interfaces: Serial0/0 Inside interfaces: Ethernet0/0

CIDR, Private Addresses, and NAT

Example 4-2

133

Dynamic NAT Configuration (Continued)

Hits: 0

Misses: 0

Expired translations: 0 Dynamic mappings: -- Inside Source access-list 1 pool fred refcount 0 pool fred: netmask 255.255.255.252 start 200.1.1.1 end 200.1.1.2 type generic, total addresses 2, allocated 0 (0%), misses 0 ! At this point, a Telnet session from 10.1.1.1 to 170.1.1.1 started. ! Below, the 1 “miss” means that the first packet from 10.1.1.2 did not have a ! matching entry in the table, but that packet triggered NAT to add an entry to the ! NAT table. Host 10.1.1.2 has then sent 69 more packets, noted as “hits” because ! there was an entry in the table. NAT# show ip nat statistics Total active translations: 1 (0 static, 1 dynamic; 0 extended) Outside interfaces: Serial0/0 Inside interfaces: Ethernet0/0 Hits: 69

Misses: 1

Expired translations: 0 Dynamic mappings: -- Inside Source access-list 1 pool fred refcount 1 pool fred: netmask 255.255.255.252 start 200.1.1.1 end 200.1.1.2 type generic, total addresses 2, allocated 1 (50%), misses 0 ! The dynamic NAT entry is now displayed in the table. NAT# show ip nat translations Pro Inside global

Inside local

Outside local

Outside global

--- 200.1.1.1

10.1.1.2

---

---

! Below, the configuration uses PAT via the overload parameter. Could have used the ! ip nat inside source list 1 int s0/0 overload command instead, using a single ! IP Inside Global IP address. NAT(config)# no ip nat inside source list 1 pool fred NAT(config)# ip nat inside source list 1 pool fred overload ! To test, the dynamic NAT entries were cleared after changing the NAT ! configuration. Before the next command was issued, host 10.1.1.1 had created two ! Telnet connections, and host 10.1.1.2 created 1 more TCP connection. NAT# clear ip nat translations * ! Before the next command was issued, host 10.1.1.1 had created two ! Telnet connections, and host 10.1.1.2 created 1 more TCP connection. Note that ! all three dynamically mapped flows use common Inside Global 200.1.1.1.

continues

134

Chapter 4: IP Addressing

Example 4-2

Dynamic NAT Configuration (Continued)

NAT# show ip nat translations Pro Inside global

Inside local

Outside local

Outside global

tcp 200.1.1.1:3212

10.1.1.1:3212

170.1.1.1:23

170.1.1.1:23

tcp 200.1.1.1:3213

10.1.1.1:3213

170.1.1.1:23

170.1.1.1:23

tcp 200.1.1.1:38913

10.1.1.2:38913

170.1.1.1:23

170.1.1.1:23

Foundation Summary

135

Foundation Summary This section lists additional details and facts to round out the coverage of the topics in this chapter. Unlike most of the Cisco Press Exam Certification Guides, this “Foundation Summary” does not repeat information presented in the “Foundation Topics” section of the chapter. Please take the time to read and study the details in the “Foundation Topics” section of the chapter, as well as review items noted with a Key Topic icon. Table 4-14 lists and briefly explains several variations on NAT. Table 4-14

Variations on NAT Name

Function

Static NAT

Statically correlates the same public IP address for use by the same local host every time. Does not conserve IP addresses.

Dynamic NAT

Pools the available public IP addresses, shared among a group of local hosts, but with only one local host at a time using a public IP address. Does not conserve IP addresses.

Dynamic NAT with overload (PAT)

Like dynamic NAT, but multiple local hosts share a single public IP address by multiplexing using TCP and UDP port numbers. Conserves IP addresses.

NAT for overlapping address

Can be done with any of the first three types. Translates both source and destination addresses, instead of just the source (for packets going from enterprise to the Internet).

Table 4-15 lists the protocols mentioned in this chapter and their respective standards documents. Table 4-15

Protocols and Standards for Chapter 4 Name

Standardized In

IP

RFC 791

Subnetting

RFC 950

NAT

RFC 1631

Private addressing

RFC 1918

CIDR

RFCs 1517–1520

136

Chapter 4: IP Addressing

Table 4-16 lists and describes some of the most commonly used IOS commands related to the topics in this chapter. Table 4-16

Command Reference for Chapter 4 Command

Description

ip address ip-address mask [secondary]

Interface subcommand to assign an IPv4 address

ip nat {inside | outside}

Interface subcommand; identifies inside or outside part of network

ip nat inside source {list {access-list-number | access-list-name} | route-map name} {interface type number | pool pool-name} [overload]

Global command that defines the set of inside addresses for which NAT will be performed, and corresponding outside addresses

ip nat inside destination list {access-listnumber | name} pool name

Global command used with destination NAT

ip nat outside source {list {access-list-number | access-list-name} | route-map name} pool poolname [add-route]

Global command used with both destination and dynamic NAT

ip nat pool name start-ip end-ip {netmask netmask | prefix-length prefix-length}[type rotary]

Global command to create a pool of addresses for dynamic NAT

show ip nat statistics

Lists counters for packets and for NAT table entries, as well as basic configuration information

show ip nat translations [verbose]

Displays the NAT table

clear ip nat translation {* | [inside global-ip local-ip] [outside local-ip global-ip]}

Clears all or some of the dynamic entries in the NAT table, depending on which parameters are used

debug ip nat

Issues log messages describing each packet whose IP address is translated with NAT

show ip interface [type number] [brief]

Lists information about IPv4 on interfaces

Foundation Summary

137

Figure 4-8 shows the IP header format. Figure 4-8

IP Header 0

8 Version

Header Length

16 DS Field

Identification Time to Live

24

32

Packet Length Flags (3)

Protocol

Fragment Offset (13) Header Checksum

Source IP Address Destination IP Address Optional Header Fields and Padding

Table 4-17 lists the terms and meanings of the fields inside the IP header. Table 4-17

IP Header Fields Field

Meaning

Version

Version of the IP protocol. Most networks use IPv4 today, with IPv6 becoming more popular. The header format reflects IPv4.

Header Length

Defines the length of the IP header, including optional fields. Because the length of the IP header must always be a multiple of 4, the IP header length (IHL) is multiplied by 4 to give the actual number of bytes.

DS Field

Differentiated Services Field. This byte was originally called the Type of Service (ToS) byte, but was redefined by RFC 2474 as the DS Field. It is used for marking packets for the purpose of applying different quality of service (QoS) levels to different packets.

Packet Length

Identifies the entire length of the IP packet, including the data.

Identification

Used by the IP packet fragmentation process. If a single packet is fragmented into multiple packets, all fragments of the original packet contain the same identifier, so that the original packet can be reassembled.

Flags

3 bits used by the IP packet fragmentation process.

Fragment Offset

A number set in a fragment of a larger packet that identifies the fragment’s location in the larger original packet.

Time to Live (TTL)

A value used to prevent routing loops. Routers decrement this field by 1 each time the packet is forwarded; once it decrements to 0, the packet is discarded. continues

138

Chapter 4: IP Addressing

Table 4-17

IP Header Fields (Continued) Field

Meaning

Protocol

A field that identifies the contents of the data portion of the IP packet. For example, protocol 6 implies a TCP header is the first thing in the IP packet data field.

Header Checksum

A value used to store a frame check sequence (FCS) value, whose purpose is to determine if any bit errors occurred in the IP header (not the data) during transmission.

Source IP Address

The 32-bit IP address of the sender of the packet.

Destination IP Address

The 32-bit IP address of the intended recipient of the packet.

Optional Header Fields and Padding

IP supports additional header fields for future expansion via optional headers. Also, if these optional headers do not use a multiple of 4 bytes, padding bytes are added, comprised of all binary 0s, so that the header is a multiple of 4 bytes in length.

Table 4-18 lists some of the more common IP protocol field values. Table 4-18

IP Protocol Field Values Protocol Name

Protocol Number

ICMP

1

TCP

6

UDP

17

EIGRP

88

OSPF

89

PIM

103

Memory Builders The CCIE Routing and Switching written exam, like all Cisco CCIE written exams, covers a fairly broad set of topics. This section provides some basic tools to help you exercise your memory about some of the broader topics covered in this chapter.

Fill in Key Tables from Memory Appendix G, “Key Tables for CCIE Study,” on the CD in the back of this book contains empty sets of some of the key summary tables in each chapter. Print Appendix G, refer to this chapter’s tables

Memory Builders

139

in it, and fill in the tables from memory. Refer to Appendix H, “Solutions for Key Tables for CCIE Study,” on the CD to check your answers.

Definitions Next, take a few moments to write down the definitions for the following terms: subnet, prefix, classless IP addressing, classful IP addressing, CIDR, NAT, IPv4, subnet broadcast address, subnet number, subnet zero, broadcast subnet, subnet mask, private addresses, SLSM, VLSM, Inside Local address, Inside Global address, Outside Local address, Outside Global address, PAT, overloading, quartet Refer to the glossary to check your answers.

Further Reading All topics in this chapter are covered to varying depth for the CCNP Routing exam. For more details on these topics, look for the CCNP routing study guides at www.ciscopress.com/ccnp.

Blueprint topics covered in this chapter: This chapter covers the following subtopics from the Cisco CCIE Routing and Switching written exam blueprint. Refer to the full blueprint in Table I-1 in the Introduction for more details on the topics covered in each chapter and their context within the blueprint. ■ ■

Hot Standby Router Protocol (HSRP) Gateway Load Balancing Protocol (GLBP)

■

Virtual Router Redundancy Protocol (VRRP)

■

Dynamic Host Configuration Protocol (DHCP)

■

Network Time Protocol (NTP)

■

Web Cache Communication Protocol (WCCP)

■

Network Management

■

Logging and Syslog

■

Troubleshoot Network Services

■

Implement IP Service Level Agreement (IP SLA)

■

Implement NetFlow

■

Implement Router IP Traffic Export (RITE)

■

Implement SNMP

■

Implement Cisco IOS Embedded Event Manager (EEM)

■

Implement Remote Monitoring (RMON)

■

Implement FTP

■

Implement TFTP

■

Implement TFTP Server on Router

■

Implement Secure Copy Protocol (SCP)

■

Implement HTTP and HTTPS

■

Implement Telnet

■

Implement SSH

CHAPTER

5

IP Services IP relies on several protocols to perform a variety of tasks related to the process of routing packets. This chapter provides a reference for the most popular of these protocols. In addition, this chapter covers a number of management-related protocols and other blueprint topics related to IP services.

“Do I Know This Already?” Quiz Table 5-1 outlines the major headings in this chapter and the corresponding “Do I Know This Already?” quiz questions. Table 5-1

“Do I Know This Already?” Foundation Topics Section-to-Question Mapping Foundation Topics Section

Questions Covered in This Section

ARP, Proxy ARP, Reverse ARP, BOOTP, and DHCP

1–3

HSRP, VRRP, and GLBP

4–6

Network Time Protocol

7

SNMP

8–9

Web Cache Communication Protocol

10–11

Implement SSH

12

Implement SSH, HTTPS, FTP, SCP, TFTP

13

Implement RMON

14

Implement IP SLA, NetFlow, RITE, EEM

15

Score

Total Score

To best use this pre-chapter assessment, remember to score yourself strictly. You can find the answers in Appendix A, “Answers to the ‘Do I Know This Already?’ Quizzes.” 1.

Two hosts, named PC1 and PC2, sit on subnet 172.16.1.0/24, along with router R1. A web server sits on subnet 172.16.2.0/24, which is connected to another interface of R1. At some point, both PC1 and PC2 send an ARP request before they successfully send packets to the

142

Chapter 5: IP Services

web server. With PC1, R1 makes a normal ARP reply, but for PC2, R1 uses a proxy ARP reply. Which two of the following answers could be true given the stated behavior in this network?

2.

3.

4.

a.

PC2 set the proxy flag in the ARP request.

b.

PC2 encapsulated the ARP request inside an IP packet.

c.

PC2’s ARP broadcast implied that PC2 was looking for the web server’s MAC address.

d.

PC2 has a subnet mask of 255.255.0.0.

e.

R1’s proxy ARP reply contains the web server’s MAC address.

Host PC3 is using DHCP to discover its IP address. Only one router attaches to PC3’s subnet, using its fa0/0 interface, with an ip helper-address 10.5.5.5 command on that same interface. That same router interface has an ip address 10.4.5.6 255.255.252.0 command configured as well. Which of the following are true about PC3’s DHCP request? a.

The destination IP address of the DHCP request packet is set to 10.5.5.5 by the router.

b.

The DHCP request packet’s source IP address is unchanged by the router.

c.

The DHCP request is encapsulated inside a new IP packet, with source IP address 10.4.5.6 and destination 10.5.5.5.

d.

The DHCP request’s source IP address is changed to 10.4.5.255.

e.

The DHCP request’s source IP address is changed to 10.4.7.255.

Which of the following statements are true about BOOTP, but not true about RARP? a.

The client can be assigned a different IP address on different occasions, because the server can allocate a pool of IP addresses for allocation to a set of clients.

b.

The server can be on a different subnet from the client.

c.

The client’s MAC address must be configured on the server, with a one-to-one mapping to the IP address to be assigned to the client with that MAC address.

d.

The client can discover its IP address, subnet mask, and default gateway IP address.

R1 is HSRP active for virtual IP address 172.16.1.1, with HSRP priority set to 115. R1 is tracking three separate interfaces. An engineer configures the same HSRP group on R2, also connected to the same subnet, only using the standby 1 ip 172.16.1.1 command, and no other HSRP-related commands. Which of the following would cause R2 to take over as HSRP active? a.

R1 experiences failures on tracked interfaces, totaling 16 or more lost points.

b.

R1 experiences failures on tracked interfaces, totaling 15 or more lost points.

c.

R2 could configure a priority of 116 or greater.

d.

R1’s fa0/0 interface fails.

e.

R2 would take over immediately.

“Do I Know This Already?” Quiz

5.

6.

7.

8.

9.

143

Which Cisco IOS feature does HSRP, GLBP, and VRRP use to determine when an interface fails for active switching purposes? a.

Each protocol has a built-in method of tracking interfaces.

b.

When a physical interface goes down, the redundancy protocol uses this automatically as a basis for switching.

c.

Each protocol uses its own hello mechanism for determining which interfaces are up or down.

d.

The Cisco IOS object tracking feature.

Which is the correct term for using more than one HSRP group to provide load balancing for HSRP? a.

LBHSRP

b.

LSHSRP

c.

RHSRP

d.

MHSRP

e.

None of these. HSRP does not support load balancing.

Which of the following NTP modes in a Cisco router requires a predefinition of the IP address of an NTP server? a.

Server mode

b.

Static client mode

c.

Broadcast client mode

d.

Symmetric active mode

Which of the following are true about SNMP security? a.

SNMP Version 1 calls for the use of community strings that are passed as clear text.

b.

SNMP Version 2c calls for the use of community strings that are passed as MD5 message digests generated with private keys.

c.

SNMP Version 3 allows for authentication using MD5 message digests generated with private keys.

d.

SNMP Version 3 authentication also requires concurrent use of encryption, typically done with DES.

Which of the following statements are true regarding features of SNMP based on the SNMP version? a.

SNMP Version 2 added the GetNext protocol message to SNMP.

b.

SNMP Version 3 added the Inform protocol message to SNMP.

144

Chapter 5: IP Services

10.

11.

12.

13.

c.

SNMP Version 2 added the Inform protocol message to SNMP.

d.

SNMP Version 3 expanded the SNMP Response protocol message so that it must be used by managers in response to Traps sent by agents.

e.

SNMP Version 3 enhanced SNMP Version 2 security features but not other features.

WCCP uses what protocol and port for communication between content engines and WCCP routers? a.

UDP 2048

b.

TCP 2048

c.

UDP 4082

d.

TCP 4082

In a WCCP cluster, which content engine becomes the lead engine after the cluster stabilizes? a.

The content engine with the lowest IP address.

b.

The content engine with the highest IP address.

c.

There is no such thing as a lead content engine; the correct term is designated content engine.

d.

All content engines have equal precedence for redundancy and the fastest possible load sharing.

Which configuration commands are required to enable SSH on a router? a.

hostname

b.

ip domain-name

c.

ip ssh

d.

crypto key generate rsa

e.

http secure-server

Which protocol is the most secure choice, natively, for transferring files from a router? a.

SSH

b.

HTTPS

c.

FTP

d.

TFTP

e.

SCP

“Do I Know This Already?” Quiz

14.

15.

In RMON, which type of configured option includes rising and falling thresholds, either relative or absolute, and is monitored by another type of RMON option? a.

Event

b.

Alert

c.

Notification

d.

Port

e.

Probe

Which Cisco IOS feature permits end-to-end network performance monitoring with configuration on devices at each end of the network? a.

Flexible NetFlow

b.

IP SLA

c.

EEM

d.

RITE

145

146

Chapter 5: IP Services

Foundation Topics

ARP, Proxy ARP, Reverse ARP, BOOTP, and DHCP The heading for this section may seem like a laundry list of a lot of different protocols. However, these five protocols do have one central theme, namely that they help a host learn information so that it can successfully send and receive IP packets. Specifically, ARP and proxy ARP define methods for a host to learn another host’s MAC address, whereas the core functions of RARP, BOOTP, and DHCP define how a host can discover its own IP address, plus additional related information.

ARP and Proxy ARP You would imagine that anyone getting this far in their CCIE study would already have a solid understanding of the Address Resolution Protocol (ARP, RFC 826). However, proxy ARP (RFC 1027) is often ignored, in part because of its lack of use today. To see how they both work, Figure 5-1 shows an example of each, with Fred and Barney both trying to reach the web server at IP address 10.1.2.200. Figure 5-1

Comparing ARP and Proxy ARP

Fred: 10.1.1.101 Mask: /24 GW = 10.1.1.1

ARP Request Target = 10.1.1.1 ARP Reply Source = R1-E1-MAC

Web Server: 10.1.2.200/24 GW = 10.1.2.1

Interface: E1 MAC: R1-E1-MAC 10.1.1.1/24

R1 Interface: E2 MAC: R1-E2-MAC 10.1.2.1/24 Barney: 10.1.1.102 Mask: /8 GW = 10.1.1.1

ARP Request Target = 10.1.2.200 ARP Reply Source = R1-E1-MAC

DHCP Server: 10.1.2.202/24

DNS Server: 10.1.2.203/24

Proxy logic used by R1!

Fred follows a normal ARP process, broadcasting an ARP request, with R1’s E1 IP address as the target. The ARP message has a Target field of all 0s for the MAC address that needs to be learned, and a target IP address of the IP address whose MAC address it is searching, namely 10.1.1.1 in

ARP, Proxy ARP, Reverse ARP, BOOTP, and DHCP

147

this case. The ARP reply lists the MAC address associated with the IP address, in this case, the MAC address of R1’s E1 interface. NOTE The ARP message itself does not include an IP header, although it does have destination and source IP addresses in the same relative position as an IP header. The ARP request lists an IP destination of 255.255.255.255. The ARP Ethernet protocol type is 0x0806, whereas IP packets have an Ethernet protocol type of 0x0800.

Proxy ARP uses the exact same ARP message as ARP, but the ARP request is actually requesting a MAC address that is not on the local subnet. Because the ARP request is broadcast on the local subnet, it will not be heard by the target host—so if a router can route packets to that target host, the router issues a proxy ARP reply on behalf of that target. For instance, Barney places the web server’s IP address (10.1.2.200) in the Target field, because Barney thinks that he is on the same subnet as the web server due to Barney’s mask of 255.0.0.0. The ARP request is a LAN broadcast, so R1, being a well-behaved router, does not forward the ARP broadcast. However, knowing that the ARP request will never get to the subnet where 10.1.2.200 resides, R1 saves the day by replying to the ARP on behalf of the web server. R1 takes the web server’s place in the ARP process, hence the name proxy ARP. Also, note that R1’s ARP reply contains R1’s E1 MAC address, so that Barney will forward frames to R1 when Barney wants to send a packet to the web server. Before the advent of DHCP, many networks relied on proxy ARP, configuring hosts to use the default masks in their respective networks. Regardless of whether the proxy version is used, the end result is that the host learns a router’s MAC address to forward packets to another subnet.

RARP, BOOTP, and DHCP The ARP and proxy ARP processes both occur after a host knows its IP address and subnet mask. RARP, BOOTP, and DHCP represent the evolution of protocols defined to help a host dynamically learn its IP address. All three protocols require the client host to send a broadcast to begin discovery, and all three rely on a server to hear the request and supply an IP address to the client. Figure 5-2 shows the basic processes with RARP and BOOTP.

148

Chapter 5: IP Services

Figure 5-2

RARP and BOOTP—Basic Processes

RARP 1 RARP Broadcast

Configuration RARP Server

Hannah

MAC 0200.1111.1111 0200.1234.5678 0200.5432.1111

2 RARP Reply IP: ?.?.?.? MAC: 0200.1111.1111

1

IP 10.1.1.1 10.1.1.2 10.1.1.3

2 Hey Everybody! My MAC Address Is 0200.1111.1111. If You Are a RARP Server, Please Tell Me My IP Address!

Your IP Address Is 10.1.1.1

BOOTP

Configuration MAC IP 0200.1111.1111 10.1.1.1 0200.1234.5678 10.1.1.2 0200.5432.1111 10.1.1.3

Hannah

1 BOOTP Broadcast

BOOTP Server

2 BOOTP Reply 1

IP: ?.?.?.? MAC: 0200.1111.1111 2

Hey Everybody! My MAC Address Is 0200.1111.1111. If You Are a BOOTP Server, Please Tell Me My IP Address!

Your IP Address Is 10.1.1.1 Your Default Gateway Is 10.1.1.200 …

Gateway 10.1.1.200 10.1.1.200 10.1.1.200

10.1.1.200

R1

A RARP request is a host’s attempt to find its own IP address. So RARP uses the same old ARP message, but the ARP request lists a MAC address target of its own MAC address and a target IP address of 0.0.0.0. A preconfigured RARP server, which must be on the same subnet as the client, receives the request and performs a table lookup in its configuration. If that target MAC address listed in the ARP request is configured on the RARP server, the RARP server sends an ARP reply, after entering the configured IP address in the Source IP address field. BOOTP was defined in part to improve IP address assignment features of RARP. BOOTP uses a completely different set of messages, defined by RFC 951, with the commands encapsulated inside an IP and UDP header. With the correct router configuration, a router can forward the BOOTP packets to other subnets—allowing the deployment of a centrally located BOOTP server. Also, BOOTP supports the assignment of many other tidbits of information, including the subnet mask, default gateway, DNS addresses, and its namesake, the IP address of a boot (or image) server. However, BOOTP does not solve the configuration burden of RARP, still requiring that the server be preconfigured with the MAC addresses and IP addresses of each client.

DHCP DHCP represents the next step in the evolution of dynamic IP address assignment. Building on the format of BOOTP protocols, DHCP focuses on dynamically assigning a variety of information and provides flexible messaging to allow for future changes, without requiring predefinition of MAC addresses for each client. DHCP also includes temporary leasing of IP addresses, enabling address reclamation, pooling of IP addresses, and, recently, dynamic registration of client DNS

ARP, Proxy ARP, Reverse ARP, BOOTP, and DHCP

149

fully qualified domain names (FQDNs). (See http://www.ietf.org for more information on FQDN registration.) DHCP servers typically reside in a centralized location, with remote routers forwarding the LANbroadcast DHCP requests to the DHCP server by changing the request’s destination address to match the DHCP server. This feature is called DHCP relay agent. For instance, in Figure 5-1, if Fred and Barney were to use DHCP, with the DHCP server at 10.1.2.202, R1 would change Fred’s DHCP request from a destination of 255.255.255.255 to a destination of 10.1.2.202. R1 would also list its own IP address in the message, in the gateway IP address (giaddr) field, notifying the DHCP server the IP address to which the response should be sent. After receiving the next DHCP message from the server, R1 would change the destination IP address to a LAN broadcast, and forward the packet onto the client’s LAN. The only configuration requirement on the router is an ip helper-address 10.1.2.202 interface subcommand on its E1 interface. Alternatively, R1 could be configured as a DHCP server—a feature that is not often configured on routers in production networks but is certainly fair game for the CCIE written and lab exams. Configuring DHCP on a router consists of several required steps: Step 1

Configure a DHCP pool.

Step 2

Configure the router to exclude its own IP address from the DHCP pool.

Step 3

Disable DHCP conflict logging or configure a DHCP database agent.

The DHCP pool includes key items such as the subnet (using the network command within DHCP pool configuration), default gateway (default-router), and the length of time for which the DHCP lease is valid (lease). Other items, including the DNS domain name and any DHCP options, are also defined within the DHCP pool. Although not strictly necessary in DHCP configuration, it is certainly a best practice to configure the router to make its own IP address in the DHCP pool subnet unavailable for allocation via DHCP. The same is true for any other static IP addresses within the DHCP pool range, such as those of servers and other routers. Exclude host IP addresses from the DHCP process using the ip dhcp excluded-address command. NOTE The ip dhcp excluded-address command is one of the relatively few Cisco IOS ip commands that is a global configuration command rather than an interface command.

The Cisco IOS DHCP server also provides a mechanism for logging DHCP address conflicts to a central server called a DHCP database agent. IOS requires that you either disable conflict logging by using the no ip dhcp conflict-logging command or configure a DHCP database agent on a

150

Chapter 5: IP Services

server by using the ip dhcp database command. Example 5-1 shows R1’s configuration for a DHCP relay agent, as well as an alternative for R1 to provide DNS services for subnet 10.1.1.0/24. Example 5-1

DHCP Configuration Options—R1, Figure 5-1

! UDP broadcasts coming in E0 will be forwarded as unicasts to 10.1.2.202. ! The source IP will be changed to 10.1.1.255, so that the reply packets will be ! broadcast back out E0. interfac e Ethernet1 ip address 10.1.1.1 255.255.255.0 i p h e l p e r - a d d r e s s 1 0 . 1 . 2. 2 0 2 ! Below, an alternative configuration, with R1 as the DHCP server. R1 assigns IP ! addresses other than the excluded first 20 IP addresses in the subnet, and informs the ! clients of their IP addresses, mask, DNS, and default router. Leases are for 0 days, ! 0 hours, and 20 minutes. ip dhcp excluded-address 10.1.1.0 10.1.1.2 0 ! ip dhcp p ool subnet1 n et w o r k 1 0 . 1 . 1 . 0 2 5 5 . 2 5 5 . 2 5 5 . 0 d n s - s e r v e r 1 0 . 1 . 2 . 20 3 d e f a u l t - r o u t e r 1 0 . 1 . 1. 1 lease 0 0 20

Table 5-2 summarizes some of the key comparison points with RARP, BOOTP, and DHCP. Table 5-2

Comparing RARP, BOOTP, and DHCP Feature

RARP

BOOTP

DHCP

Relies on server to allocate IP addresses

Yes

Yes

Yes

Encapsulates messages inside IP and UDP, so they can be forwarded to a remote server

No

Yes

Yes

Client can discover its own mask, gateway, DNS, and download server

No

Yes

Yes

Dynamic address assignment from a pool of IP addresses, without requiring knowledge of client MACs

No

No

Yes

Allows temporary lease of IP address

No

No

Yes

Includes extensions for registering client’s FQDN with a DNS

No

No

Yes

HSRP, VRRP, and GLBP IP hosts can use several methods of deciding which default router or default gateway to use— DHCP, BOOTP, ICMP Router Discovery Protocol (IRDP), manual configuration, or even by running a routing protocol (although having hosts run a routing protocol is not common today). The most typical methods—using DHCP or manual configuration—result in the host knowing a

HSRP, VRRP, and GLBP

151

single IP address of its default gateway. Hot Standby Router Protocol (HSRP), Virtual Router Redundancy Protocol (VRRP), and Gateway Load Balancing Protocol (GLBP) represent a chronological list of some of the best tools for overcoming the issues related to a host knowing a single IP address as its path to get outside the subnet. HSRP allows multiple routers to share a virtual IP and MAC address so that the end-user hosts do not realize when a failure occurs. Some of the key HSRP features are as follows: ■

Virtual IP address and virtual MAC active on the Master router

■

Standby routers listen for Hellos from the Active router, defaulting to a 3-second hello interval and 10-second dead interval

■

Highest priority (IOS default 100, range 1–255) determines the Active router, with preemption disabled by default

■

Supports tracking, whereby a router’s priority is decreased when a tracked object (interface or route) fails

■

Up to 255 HSRP groups per interface, enabling an administrative form of load balancing

■

Virtual MAC of 0000.0C07.ACxx, where xx is the hex HSRP group

■

Virtual IP address must be in the same subnet as the routers’ interfaces on the same LAN

■

Virtual IP address must be different from any of routers’ individual interface IP addresses

■

Supports clear-text and MD5 authentication (through a key chain)

Example 5-2 shows a typical HSRP configuration, with two groups configured. Routers R1 and R2 are attached to the same subnet, 10.1.1.0/24, both with WAN links (S0/0.1) connecting them to the rest of an enterprise network. Cisco IOS provides the tracking mechanism shown in Example 5-2 to permit many processes, including HSRP, VRRP, and GLBP, to track interface states. A tracking object can track based on the line protocol (shown here) or the IP routing table. The example contains the details and explanation of the configuration. Example 5-2

HSRP Configuration

! First, on Router R1, a tracking object must be configured so that ! HSRP can track the interface state. t r a c k 1 3 i n t e r f a c e S e r i a l0 / 0 . 1 l i n e - p r o t o c o l ! Next, on Router R1, two HSRP groups are configured. R1 has a higher priority ! in group 21, with R2 having a higher priority in group 22. R1 is set to preempt ! in group 21, as well as to track interface s0/0.1 for both groups. interfac e FastEthernet0/ 0 i p a d d r e s s 1 0 . 1 . 1 . 2 2 5 5 .2 5 5 . 2 5 5 . 0 standby 21 ip 10.1.1.21

continues

152

Chapter 5: IP Services

Example 5-2

HSRP Configuration (Continued)

standby 21 priority 105 standb y 21 preempt standby 21 track 13 standby 22 ip 10.1.1.22 standby 22 track 13 ! Next, R2 is configured with a higher priority for HSRP group 22, and with ! HSRP tracking enabled in both groups. The tracking “decrement” used by R2, ! when S0/0.1 fails, is set to 9 (instead of the default of 10). ! A tracking object must be configured first, as on R1. t r a c k 2 3 i n t e r f a c e S e r i a l0 / 0 . 1 l i n e - p r o t o c o l i n t e r fa c e F a s t E t h e r n e t 0 / 0 i p a dd r e s s 1 0 . 1 . 1 . 1 2 5 5 . 2 5 5 . 2 5 5 . 0 standb y 21 ip 10.1.1.21 stand by 21 track 23 standby 22 ip 10.1.1.22 s t an d b y 2 2 p r i o r i t y 1 0 5 standby 22 track 23 decrement 9 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! ! On R1 below, for group 21, the output shows that R1 is active, with R2 ! (10.1.1.2) as standby. ! R1 is tracking s0/0.1, with a default “decrement” of 10, meaning that the ! configured priority of 105 will be decremented by 10 if s0/0.1 fails. Router1# sh standby fa0/0 FastEthernet0/0 - Group 21 State is Active 2 state changes, last state change 00:00:45 Virtual IP address is 10.1.1.21 Active virtual MAC address is 0000.0c07.ac15 Local virtual MAC address is 0000.0c07.ac15 (v1 default) Hello time 3 sec, hold time 10 sec Next hello sent in 2.900 secs Preemption enabled Active router is local Standby router is 10.1.1.2, priority 100 (expires in 7.897 sec) Priority 105 (configured 105) Track object 13 state Up decrement 10 IP redundancy name is "hsrp-Fa0/0-21" (default) ! output omitted ! NOT SHOWN—R1 shuts down S0.0.1, lowering its priority in group 21 by 10. ! The debug below shows the reduced priority value. However, R2 does not become ! active, because R2’s configuration did not include a standby 21 preempt command. Router1# debug standby *Mar

1 00:24:04.122: HSRP: Fa0/0 Grp 21 Hello

out 10.1.1.1 Active

pri 95 vIP 10.1.1.21

HSRP, VRRP, and GLBP

153

Because HSRP uses only one Active router at a time, any other HSRP routers are idle. To provide load sharing in an HSRP configuration, the concept of Multiple HSRP, or MHSRP, was developed. In MHSRP, two or more HSRP groups are configured on each HSRP LAN interface, where the configured priority determines which router will be active for each HSRP group. MHSRP requires that each DHCP client and statically configured host is issued a default gateway corresponding to one of the HSRP groups and requires that they’re distributed appropriately. Thus, in an MHSRP configuration with two routers and two groups, all other things being equal, half of the hosts should have one HSRP group address as its default gateway, and the other half of the hosts should use the other HSRP group address. If you now revisit Example 5-2, you will see that it is an MHSRP configuration. HSRP is Cisco proprietary, has been out a long time, and is widely popular. VRRP (RFC 3768) provides a standardized protocol to perform almost the exact same function. The Cisco VRRP implementation has the same goals in mind as HSRP but with these differences: ■

VRRP uses a multicast virtual MAC address (0000.5E00.01xx, where xx is the hex VRRP group number).

■

VRRP uses the IOS object tracking feature, rather than its own internal tracking mechanism, to track interface states for failover purposes.

■

VRRP defaults to use pre-emption, but HSRP defaults to not use pre-emption. Both can be configured to either use pre-emption or not.

■

The VRRP term Master means the same thing as the HSRP term Active.

■

In VRRP, the VRRP group IP address is the interface IP address of one of the VRRP routers.

GLBP is a newer Cisco-proprietary tool that adds load-balancing features in addition to gatewayredundancy features. Hosts still point to a default gateway IP address, but GLBP causes different hosts to send their traffic to one of up to four routers in a GLBP group. To do so, the GLBP Active Virtual Gateway (AVG) assigns each router in the group a unique virtual MAC address, following the format 0007.B400.xxyy, where xx is the GLBP group number, and yy is a different number for each router (01, 02, 03, or 04). When a client ARPs for the (virtual) IP address of its default gateway, the GLBP AVG replies with one of the four possible virtual MACs. By replying to ARP requests with different virtual MACs, the hosts in that subnet will in effect balance the traffic across the routers, rather than send all traffic to the one active router. Cisco IOS devices with GLBP support permit configuring up to 1024 GLBP groups per physical interface and up to four hosts per GLBP group.

154

Chapter 5: IP Services

Network Time Protocol NTP Version 3 (RFC 1305) allows IP hosts to synchronize their time-of-day clocks with a common source clock. For instance, routers and switches can synchronize their clocks to make event correlation from an SNMP management station more meaningful, by ensuring that any events and traps have accurate time stamps. By design, most routers and switches use NTP client mode, adjusting their clocks based on the time as known by an NTP server. NTP defines the messages that flow between client and server, and the algorithms a client uses to adjust its clock. Routers and switches can also be configured as NTP servers, as well as using NTP symmetric active mode—a mode in which the router or switch mutually synchronizes with another NTP host. NTP servers may reference other NTP servers to obtain a more accurate clock source as defined by the stratum level of the ultimate source clock. For instance, atomic clocks and Global Positioning System (GPS) satellite transmissions provide a source of stratum 1 (lowest/best possible stratum level). For an enterprise network, the routers and switches can refer to a lowstratum NTP source on the Internet, or purpose-built rack-mounted NTP servers, with built-in GPS capabilities, can be deployed. Example 5-3 shows a sample NTP configuration on four routers, all sharing the same 10.1.1.0/24 Ethernet subnet. Router R1 will be configured as an NTP server. R2 acts as an NTP static client by virtue of the static configuration referencing R1’s IP address. R3 acts as an NTP broadcast client by listening for R1’s NTP broadcasts on the Ethernet. Finally, R4 acts in NTP symmetric active mode, configured with the ntp peer command. Example 5-3

NTP Configuration

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! ! First, R1’s configuration, the ntp broadcast command under interface fa0/0 ! causes NTP to broadcast NTP updates on that interface. The first three of the ! four global NTP commands configure authentication; these commands are identical ! on all the routers. R1# show running-config interface FastEthernet0/0 ntp broadcast ! ntp authentication-key 1 md5 1514190900 7 ntp authenticate ntp trusted-key 1 ntp master 7 ! Below, the “127.127.7.1” notation implies that this router is the NTP clock ! source. The clock is synchronized, with stratum level 7, as configured on the ! ntp master 7 command above. R1# show ntp associations

SNMP

Example 5-3

155

NTP Configuration (Continued) address

ref clock

*~127.127.7.1

127.127.7.1

st

when

6

22

poll reach 64

377

delay

offset

disp

0.0

0.00

0.0

* master (synced), # master (unsynced), + selected, - candidate, ~ configured R1# show ntp status Clock is synchronized, stratum 7, reference is 127.127.7.1 nominal freq is 249.5901 Hz, actual freq is 249.5901 Hz, precision is 2**16 reference time is C54483CC.E26EE853 (13:49:00.884 UTC Tue Nov 16 2004) clock offset is 0.0000 msec, root delay is 0.00 msec root dispersion is 0.02 msec, peer dispersion is 0.02 msec ! R2 is configured below as an NTP static client. Note that the ntp clock-period ! command is automatically generated as part of the synchronization process, and ! should not be added to the configuration manually. R2# show run | begin ntp ntp authentication-key 1 md5 1514190900 7 ntp authenticate ntp trusted-key 1 ntp clock-period 17208144 ntp server 10.1.1.1 end ! Next, R3 has been configured as an NTP broadcast client. The ntp broadcast client ! command on R3 tells it to listen for the broadcasts from R1. This configuration ! relies on the ntp broadcast subcommand on R1’s Fa0/0 interface, as shown at the ! beginning of this example. R3# show run interface Ethernet0/0 ntp broadcast client ! R4’s configuration is listed, with the ntp peer ! command implying the use of symmetric active mode. R4# show run | beg ntp ntp authentication-key 1 md5 0002010300 7 ntp authenticate ntp trusted-key 1 ntp clock-period 17208233 ntp peer 10.1.1.1

SNMP This section of the chapter summarizes some of the core Simple Network Management Protocol (SNMP) concepts and details, particularly with regard to features of different SNMP versions. SNMP or, more formally, the Internet Standard Management Framework, uses a structure in which the device being managed (the SNMP agent) has information that the management software (the SNMP manager) wants to display to someone operating the network. Each SNMP agent keeps a database, called a Management Information Base (MIB), that holds a large variety of data about the operation of the device on which the agent resides. The manager collects the data by using SNMP.

156

Chapter 5: IP Services

SNMP has been defined with four major functional areas to support the core function of allowing managers to manage agents: ■

Data Definition—The syntax conventions for how to define the data to an agent or manager. These specifications are called the Structure of Management Information (SMI).

■

MIBs—Over 100 Internet standards define different MIBs, each for a different technology area, with countless vendor-proprietary MIBs as well. The MIB definitions conform to the appropriate SMI version.

■

Protocols—The messages used by agents and managers to exchange management data.

■

Security and Administration—Definitions for how to secure the exchange of data between agents and managers.

Interestingly, by separating SNMP into these major functional areas, each part has been improved and expanded independently over the years. However, it is important to know a few of the main features added for each official SNMP version, as well as for a pseudo-version called SNMPv2c, as summarized in Table 5-3. Table 5-3

SNMP Version Summaries SNMP Version

Description

1

Uses SMIv1, simple authentication with communities, but used MIB-I originally.

2

Uses SMIv2, removed requirement for communities, added GetBulk and Inform messages, but began with MIB-II originally.

2c

Pseudo-release (RFC 1905) that allowed SNMPv1-style communities with SNMPv2; otherwise, equivalent to SNMPv2.

3

Mostly identical to SNMPv2, but adds significantly better security, although it supports communities for backward compatibility. Uses MIB-II.

Table 5-3 hits the highlights of the comparison points between the various SNMP versions. As you might expect, each release builds on the previous one. For example, SNMPv1 defined community strings for use as simple clear-text passwords. SNMPv2 removed the requirement for community strings—however, backward compatibility for SNMP communities was defined via an optional RFC (1901). Even SNMPv3, with much better security, supports communities to allow backward compatibility. NOTE The use of SNMPv1 communities with SNMPv2, based on RFC 1901, has popularly been called SNMP Version 2c, with c referring to “communities,” although it is arguably not a legitimate full version of SNMP.

SNMP

157

The next few sections provide a bit more depth about the SNMP protocol, with additional details about some of the version differences.

SNMP Protocol Messages The SNMPv1 and SNMPv2 protocol messages (RFC 3416) define how a manager and agent, or even two managers, can communicate information. For instance, a manager can use three different messages to get MIB variable data from agents, with an SNMP Response message returned by the agent to the manager supplying the MIB data. SNMP uses UDP exclusively for transport, using the SNMP Response message to both acknowledge receipt of other protocol messages and supply SNMP information. Table 5-4 summarizes the key information about each of the SNMP protocol messages, including the SNMP version in which the message first appeared. Table 5-4

SNMP Protocol Messages (RFCs 1157 and 1905) Message

Initial Version

Response Message

Typically Sent By

Main Purpose

Get

1

Response

Manager

A request for a single variable’s value.

GetNext

1

Response

Manager

A request for the next single MIB leaf variable in the MIB tree.

GetBulk

2

Response

Manager

A request for multiple consecutive MIB variables with one request. Useful for getting complex structures, for example, an IP routing table.

Response

1

None

Agent

Used to respond with the information in Get and Set requests.

Set

1

Response

Manager

Sent by a manager to an agent to tell the agent to set a variable to a particular value. The agent replies with a Response message.

Trap

1

None

Agent

Allows agents to send unsolicited information to an SNMP manager. The manager does not reply with any SNMP message.

Inform

2

Response

Manager

A message used between SNMP managers to allow MIB data to be exchanged.

The three variations of the SNMP Get message, and the SNMP Response message, are typically used when someone is actively using an SNMP manager. When a user of the SNMP manager asks

158

Chapter 5: IP Services

for information, the manager sends one of the three types of Get commands to the agent. The agent replies with an SNMP Response message. The different variations of the Get command are useful, particularly when the manager wants to view large portions of the MIB. An agent’s entire MIB— whose structure can vary from agent to agent—can be discovered with successive GetNext requests, or with GetBulk requests, using a process called a MIB walk. The SNMP Set command allows the manager to change something on the agent. For example, the user of the management software can specify that a router interface should be shut down; the management station can then issue a Set command for a MIB variable on the agent. The agent sets the variable, which tells Cisco IOS Software to shut down the interface. SNMP Traps are unsolicited messages sent by the agent to the management station. For example, when an interface fails, a router’s SNMP agent could send a Trap to the SNMP manager. The management software could then highlight the failure information on a screen, e-mail first-level support personnel, page support, and so on. Also of note, there is no specific message in response to the receipt of a Trap; technically, of the messages in Table 5-4, only the Trap and Response messages do not expect to receive any kind of acknowledging message. Finally, the Inform message allows two SNMP managers to exchange MIB information about agents that they both manage.

SNMP MIBs SNMP Versions 1 and 2 included a standard generic MIB, with initial MIB-I (version 1, RFC 1156) and MIB-II (version 2, RFC 1213). MIB-II was actually created in between the release of SNMPv1 and v2, with SNMPv1 supporting MIB-II as well. After the creation of the MIB-II specification, the IETF SNMP working group changed the strategy for MIB definition. Instead of the SNMP working group creating standard MIBs, other working groups, in many different technology areas, were tasked with creating MIB definitions for their respective technologies. As a result, hundreds of standardized MIBs are defined. Additionally, vendors create their own vendor-proprietary MIBs. The Remote Monitoring MIB (RMON, RFC 2819) is a particularly important standardized MIB outside MIB-II. An SNMP agent that supports the RMON MIB can be programmed, through SNMP Set commands, to capture packets, calculate statistics, monitor thresholds for specific MIB variables, report back to the management station when thresholds are reached, and perform other tasks. With RMON, a network can be populated with a number of monitoring probes, with SNMP messaging used to gather the information as needed.

Syslog

159

SNMP Security SNMPv3 added solid security to the existing SNMPv2 and SNMPv2c specifications. SNMPv3 adds two main branches of security to SNMPv2: authentication and encryption. SNMPv3 specifies the use of MD5 and SHA to create a message digest for each SNMPv3 protocol message. Doing so enables authentication of endpoints and prevents data modification and masquerade types of attacks. Additionally, SNMPv3 managers and agents can use Digital Encryption Standard (DES) to encrypt the messages, providing better privacy. (SNMPv3 suggests future support of Advanced Encryption Standard [AES] as well, but that is not a part of the original SNMPv3 specifications.) The encryption feature remains separate due to the U.S. government export restrictions on DES technology. Example 5-4 shows a typical SNMP configuration with the following goals: ■

Enable SNMP and send traps to 192.168.1.100.

■

Send traps for a variety of events to the SNMP manager.

■

Set optional information to identify the router chassis, contact information, and location.

■

Set read-write access to the router from the 192.168.1.0/24 subnet (filtered by access list 33).

Example 5-4

Configuring SNMP

access-list 33 permit 192.168.1.0 0.0.0.255 snmp-server community public RW 33 snmp-server location B1 snmp-server contact [email protected] snmp-server chassis-id 2511_AccessServer_Canadice snmp-server enable traps snmp snmp-server enable traps hsrp snmp-server enable traps config snmp-server enable traps entity snmp-server enable traps bgp snmp-server enable traps rsvp snmp-server enable traps frame-relay snmp-server enable traps rtr snmp-server host 192.168.1.100 public

Syslog Event logging is nothing new to most CCIE candidates. Routers and switches, among other devices, maintain event logs that reveal a great deal about the operating conditions of that device, along with valuable time-stamp information to help troubleshoot problems or chains of events that take place.

160

Chapter 5: IP Services

By default, Cisco routers and switches do not log events to nonvolatile memory. They can be configured to do so using the logging buffered command, with an additional argument to specify the size of the log buffer. Configuring a router, for example, for SNMP management provides a means of passing critical events from the event log, as they occur, to a network management station in the form of traps. SNMP is, however, fairly involved to configure. Furthermore, if it’s not secured properly, SNMP also opens attack vectors to the device. However, disabling SNMP and watching event logs manually is at best tedious, and this approach simply does not scale. Syslog, described in RFC 5424, is a lightweight event-notification protocol that provides a middle ground between manually monitoring event logs and a full-blown SNMP implementation. It provides real-time event notification by sending messages that enter the event log to a Syslog server that you specify. Syslog uses UDP port 514 by default. Cisco IOS devices configured for Syslog, by default, send all events that enter the event log to the Syslog server. You can also configure Syslog to send only specific classes of events to the server. Syslog is a clear-text protocol that provides event notifications without requiring difficult, timeintensive configuration or opening attack vectors. In fact, it’s quite simple to configure basic Syslog operation: Step 1

Install a Syslog server on a workstation with a fixed IP address.

Step 2

Configure the logging process to send events to the Syslog server’s IP address using the logging host command.

Step 3

Configure any options, such as which severity levels (0–7) you want to send to the Syslog server using the logging trap command.

Web Cache Communication Protocol To ease pressure on congested WAN links in networks with many hosts, Cisco developed WCCP to coordinate the work of edge routers and content engines (also known as cache engines). Content engines collect frequently accessed data, usually HTTP traffic, locally, so that when hosts access the same pages the content can be delivered from the cache engine rather than crossing the WAN. WCCP differs from web proxy operation in that the hosts accessing the content have no knowledge that the content engine is involved in a given transaction. WCCP works by allowing edge routers to communicate with content engines to make each aware of the other’s presence and to permit the router to redirect traffic to the content engine as appropriate. Figure 5-3 shows how WCCP functions between a router and a content engine when a user requests a web object using HTTP.

Web Cache Communication Protocol

Figure 5-3

161

WCCP Operations Between a Router and a Content Engine HTTP Response

5

1 HTTP Get R1

HTTP Response

HTTP Get (Forwarded)

2 4A

Web Server

4B

3

Do I have the requested object?

The figure shows the following steps, with the main decision point on the content engine coming at Step 4: Step 1

The client sends an HTTP Get request with a destination address of the web server, as normal.

Step 2

The router’s WCCP function notices the HTTP Get request and redirects the packet to the content engine.

Step 3

The content engine looks at its disk storage cache to discover if the requested object is cached.

Step 4A

If the object is cached, the content engine sends an HTTP response, which includes the object, back to the client.

Step 4B

If the object is not cached, the content engine sends the original HTTP Get request on to the original server.

Step 5

If Step 4B was taken, the server replies to the client, with no knowledge that the packet was ever redirected to a content engine.

Using WCCP, which uses UDP port 2048, a router and a content engine, or a pool of content engines (known as a cluster), become aware of each other. In a cluster of content engines, the content engines also communicate with each other using WCCP. Up to 32 content engines can communicate with a single router using WCCPv1. If more than one content engine is present, the one with the lowest IP address is elected as the lead engine. WCCP also provides a means for content engines within a cluster to become aware of each other. content engines request information on the cluster members from the WCCP router, which replies

162

Chapter 5: IP Services

with a list. This permits the lead content engine to determine how traffic should be distributed to the cluster. In WCCPv1, only one router can redirect traffic to a content engine or a cluster of content engines. In WCCPv2, multiple routers and multiple content engines can be configured as a WCCP service group. This expansion permits much better scalability in content caching. Furthermore, WCCPv1 supports only HTTP traffic (TCP port 80, specifically). WCCPv2 supports several other traffic types and has other benefits compared to WCCPv1: ■

Supports TCP and UDP traffic other than TCP port 80, including FTP caching, FTP proxy handling, web caching for ports other than 80, Real Audio, video, and telephony.

■

Permits segmenting caching services provided by a caching cluster to a particular protocol or protocols, and uses a priority system for deciding which cluster to use for a particular cached protocol.

■

Supports multicast to simplify configuration.

■

Supports multiple routers (up to 32 per cluster) for redundancy and load distribution. (All content engines in a cluster must be configured to communicate with all routers in that cluster.)

■

Provides for MD5 security in WCCP communication using the global configuration command ip wccp password password.

■

Provides load distribution.

■

Supports transparent error handling.

When you enable WCCP globally on a router, the default version used is WCCPv2. Because the WCCP version is configured globally for a router, the version number affects all interfaces. However, multiple services can run on a router at the same time. Routers and content engines can also simultaneously participate in more than one service group. These WCCP settings are configured on a per-interface basis. Configuring WCCP on a router is not difficult because a lot of the configuration in a caching scenario takes place on the content engines; the routers need only minimal configuration. Example 5-5 shows a WCCPv2 configuration using MD5 authentication and multicast for WCCP communication. Example 5-5

WCCP Configuration Example

! First we enable WCCP globally on the router, ! specifying a service (web caching), a multicast address for ! the WCCP communication, and an MD5 password:

Web Cache Communication Protocol

Example 5-5

163

WCCP Configuration Example (Continued)

i p w c c p w e b - c a c h e g r o u p - a d d r e s s 2 3 9 . 1 2 8 . 1 . 1 0 0 p a s sw o r d c i s c o ! Next we configure an interface to redirect WCCP web-cache ! traffic outbound to a content engine: int fa0/ 0 i p w c c p w e b - c a ch e r e d i r e c t o u t ! Finally, inbound traffic on interface fa0/1 is excluded from redirection: int fa0/ 1

ip wccp redirect exclude in

Finally, WCCP can make use of access lists to filter traffic only for certain clients (or to exclude WCCP use for certain clients) using the ip wccp web-cache redirect-list access-list global command. WCCP can also use ACLs to determine which types of redirected traffic the router should accept from content engines, using the global command ip wccp web-cache group-list access-list.

Implementing the Cisco IOS IP Service Level Agreement (IP SLA) Feature The Cisco IOS IP SLA feature, formerly known as the Service Assurance Agent (SAA), and prior to that simply the Response Time Reporter (RTR) feature, is designed to provide a means of actively probing a network to gather performance information from it. Whereas most of the tools described in the following sections are designed to monitor and collect information, IP SLA is based on the concept of generating traffic at a specified interval, with specifically configured options, and measuring the results. It is built around a source-responder model, where one device (the source) generates traffic and either waits for a response from another device (the responder) or another device configured as a responder captures the sender’s traffic and does something with it. This model provides the ability to analyze actual network performance over time, under very specific conditions, to measure performance, avert outages, evaluate quality of service (QoS) performance, identify problems, verify SLAs, and reduce network outages. The IP SLA feature is extensively documented at http://www.cisco.com/go/ipsla. The IP SLA feature allows measuring the following parameters in network performance: ■

Delay (one way and round trip)

■

Jitter (directional)

■

Packet loss (directional)

■

Packet sequencing

■

Path (per hop)

164

Chapter 5: IP Services

■

Connectivity (through the UDP Echo, ICMP Echo, ICMP Path Echo, and TCP Connect functions)

■

Server or website download time

■

Voice-quality metrics (MOS)

Implementing the IP SLA feature requires these steps: Step 1

Configure the SLA operation type, including any required options.

Step 2

Configure any desired threshold conditions.

Step 3

Configure the responder(s), if appropriate.

Step 4

Schedule or start the operation and monitor the results for a sufficient period of time to meet your requirements.

Step 5

Review and interpret the results. You can use the Cisco IOS CLI or an SNMP manager to do this.

After IP SLA monitors have been configured, they cannot be edited or modified. You must delete an existing IP SLA monitor to reconfigure any of its options. Also, when you delete an IP SLA monitor to reconfigure it, the associated schedule for that IP SLA monitor is deleted, too. IP SLAs can use MD5 authentication. These are configured using the ip sla key-chain command. Example 5-6 shows a basic IP SLA configuration with the UDP Echo function. On the responding router, the only required command is global config ip sla monitor responder. On the originating router, the configuration shown in the example sets the source router to send UDP echo packets every 5 seconds for one day to 200.1.200.9 on port 1330. Example 5-6

IP SLA Basic Configuration

SLAdemo# config term SLAdemo(config)# ip sla monitor 1 SLAdemo(config-sla-monitor)# type udpEcho dest-ipaddr 200.1.200.9 dest-port 1330 SLAdemo(config-sla-monitor)# frequency 5 SLAdemo(config-sla-monitor)# exit SLAdemo(config)# ip sla monitor schedule 1 life 86400 start-time now

A number of show commands come in handy in verifying IP SLA performance. On the source router, the most useful commands are show ip sla monitor statistics and show ip sla monitor

Web Cache Communication Protocol

165

configuration. Here’s a sample of the show ip sla monitor statistics command for the sending router in the configuration in Example 5-6: SLAdemo# show ip sla monitor statistics Round trip time (RTT) Index 1 Latest RTT: 26 ms Latest operation start time: 19:42:44.799 EDT Tue Jun 9 2009 Latest operation return code: OK Number of successes: 228 Number of failures: 0 Operation time to live: 78863 sec

Implementing NetFlow NetFlow is a software feature set in Cisco IOS that is designed to provide network administrators information about what is happening in the network, so that those responsible for the network can make appropriate design and configuration changes and monitor for network attacks. NetFlow has been included in Cisco IOS for a long time, and has evolved through several versions (currently version 9). Cisco has renamed the feature Cisco Flexible NetFlow. It is more than just a renaming, however. The original NetFlow implementation included a fixed, seven tuple that identified a flow. Flexible NetFlow allows a user to configure the number of tuples to more specifically target a particular flow to monitor. The components of NetFlow are ■

Records—A set of predefined and user-defined key fields (such as source IP address, destination IP address, source port, and so on) for network monitoring.

■

Flow monitors—Applied to an interface, flow monitors include records, a cache, and optionally a flow exporter. The flow monitor cache collects information about flows.

■

Flow exporters—These export the cached flow information to outside systems (typically a server running a NetFlow collector).

■

Flow samplers—Designed to reduce the load on NetFlow-enabled devices, flow samplers allow specifying the sample size of traffic NetFlow analyzes to a ratio of 1:2 through 1:32768 packets. That is, the number of packets analyzed is configurable from 1/2 to 1/32768 of the packets flowing across the interface.

Configuring NetFlow in its most basic form uses predefined flow records, configured for collection by a flow monitor, and at least one flow exporter. Example 5-7 shows a basic NetFlow configuration for collecting information and statistics on IPv4 traffic using the predefined IPv4 record, and for configuring some timer settings to show their structure. An exporter is configured to send the collected information to a server at 192.168.1.110 on UDP port 1333, and with a DSCP

166

Chapter 5: IP Services

of 8 on the exported packets. The process consists of three steps: configuring the NetFlow monitor, applying it to an interface, and configuring an exporter. Example 5-7

Basic NetFlow Monitor and Exporter Configuration

EastEdge# show run

|

begin flow

flow exporter ipv4flowexport destination 192.168.1.110 dscp 8 transport udp 1333 ! flow monitor ipv4flow description Monitors all IPv4 traffic record netflow ipv4 original-input cache timeout inactive 600 cache timeout active 180 cache entries 5000 statistics packet protocol ! interface FastEthernet0/0 ip address 192.168.39.9 255.255.255.0 ip flow monitor ipv4flow input ! output omitted

You can verify NetFlow configuration using these commands: ■

show flow record

■

show flow monitor

■

show flow exporter

■

show flow interface

Implementing Router IP Traffic Export IP Traffic Export, or Router IP Traffic Export (RITE), exports IP packets to a VLAN or LAN interface for analysis. RITE does this only for traffic received on multiple WAN or LAN interfaces simultaneously as would typically take place only if the device were being targeted in a denial-of-service attack. The primary application for RITE is in intrusion detection (IDS) implementations, where duplicated traffic may indicate an attack on the network or device. In case of actual attacks where identical traffic is received simultaneously on multiple ports of a router, it’s useful to have the router send that traffic to an IDS for alerting and analysis–that’s what RITE does. When configuring RITE, you enable it and configure it to direct copied packets to the MAC address of the IDS host or protocol analyzer. You can configure forwarding of inbound traffic (the

Web Cache Communication Protocol

167

default), outbound traffic, or both, and filtering on the number of packets forwarded. Filtering can be performed with access lists and based on one-in-n packets. In Example 5-8, a router is configured with a RITE profile that’s applied to the fa0/0 interface and exports traffic to a host with the MAC address 0018.0fad.df30. The router is configured for bidirectional RITE, and to send one in every 20 inbound packets and one in every 100 outbound packets to this MAC address. The egress interface (toward the IDS host) is fa0/1. For simplicity, Example 5-8 shows only one ingress interface. Configuration for other ingress interfaces uses the same steps shown here for the fa0/0 interface. Example 5-8

Router IP Traffic Export Example

Edge# config term Edge(config)# ip traffic-export profile export-this Edge(config-rite)# interface fa0/0 Edge(config-rite)# bidirectional Edge(config-rite)# mac-address 0018.0fad.df30 Edge(config-rite)# incoming sample one-in-every 20 Edge(config-rite)# outgoing sample one-in-every 100 Edge(config-rite)# exit Edge(config)# interface fa0/1 Edge(config-if)# ip traffic-export apply export-this Edge(config-if)# end Edge# %RITE-5-ACTIVATE: Activated IP traffic export on interface FastEthernet 0/1.

Implementing Cisco IOS Embedded Event Manager The Embedded Event Manager is a software component of Cisco IOS that is designed to make life easier for administrators by tracking and classifying events that take place on a router and providing notification options for those events. Cisco’s motivation for including EEM was to reduce downtime, thus improving availability, by reducing the mean time to recover from various system events that previously required a manual troubleshooting and remediation process. In some ways, EEM overlaps with RMON functionality, but EEM is considerably more powerful and flexible. EEM uses event detectors and actions to provide notifications of those events. Event detectors that EEM supports include the following: ■

Monitoring SNMP objects

■

Screening Syslog messages for a pattern match (using regular expressions)

■

Monitoring counters

■

Timers (absolute time-of-day, countdown, watchdog, and CRON)

■

Screening CLI input for a regular expression match

168

Chapter 5: IP Services

■

Hardware insertion and removal

■

Routing table changes

■

IP SLA and NetFlow events

■

Generic On-Line Diagnostics (GOLD) events

■

Many others, including redundant switchover events, inbound SNMP messages, and others

Event actions that EEM provides include the following: ■

Generating prioritized Syslog messages

■

Reloading the router

■

Switching to a secondary processor in a redundant platform

■

Generating SNMP traps

■

Setting or modifying a counter

■

Executing a Cisco IOS command

■

Sending a brief email message

■

Requesting system information when an event occurs

■

Reading or setting the state of a tracked object

EEM policies can be written using either the Cisco IOS CLI or using the Tcl command interpreter language. For the purposes of the CCIE Routing and Switching qualification exam, you’re more likely to encounter CLI-related configuration than Tcl, but both are very well documented at http://www.cisco.com/go/eem. Here’s a brief example configuration that shows the CLI configuration of an EEM event that detects and then sends a notification that a console user has issued the wr command, as well as the associated console output when the command is issued. Example 5-9

EEM Configuration Example

R9(config)# event manager applet CLI-cp-run-st R9(config-applet)# event cli pattern "wr" sync yes R9(config-applet)# action 1.0 syslog msg "$_cli_msg Command Executed" R9(config-applet)# set 2.0 _exit_status 1 R9(config-applet)# end R9# wr Jun

9 19:23:21.989: %HA_EM-6-LOG: CLI-cp-run-st: write Command Executed

Web Cache Communication Protocol

169

The Cisco IOS EEM has such vast capability that an entire book on the subject is easily conceivable, but considering the scope of the CCIE Routing and Switching qualifying exam, these fundamental concepts should provide you with enough working knowledge to interpret questions you may encounter.

Implementing Remote Monitoring Remote Monitoring, or RMON, is an event-notification extension of the SNMP capability on a Cisco router or switch. RMON enables you to configure thresholds for alerting based on SNMP objects, so that you can monitor device performance and take appropriate action to any deviations from the normal range of performance indications. RMON is divided into two classes: alarms and events. An event is a numbered, user-configured threshold for a particular SNMP object. You configure events to track, for example, CPU utilization or errors on a particular interface, or anything else you can do with an SNMP object. You set the rising and falling thresholds for these events, and then tell RMON which RMON alarm to trigger when those rising or falling thresholds are crossed. For example, you might want to have the router watch CPU utilization and trigger an SNMP trap or log an event when the CPU utilization rises faster than, say, 20 percent per minute. Or you may configure it to trigger an alarm when the CPU utilization rises to some absolute level, such as 80 percent. Both types of thresholds (relative, or “delta,” and absolute) are supported. Then, you can configure a different alarm notification as the CPU utilization falls, again at some delta or to an absolute level you specify. The alarm that corresponds to each event is also configurable in terms of what it does (logs the event or sends a trap). If you configure an RMON alarm to send a trap, you also need to supply the SNMP community string for the SNMP server. Event and alarm numbering are locally significant. Alarm numbering provides a pointer to the corresponding event. That is, the configured events each point to specific alarm numbers, which you must also define. Here is an example of the configuration required to identify two pairs of events, and the four corresponding alarm notifications. The events being monitored are the interface error counter on the FastEthernet 0/0 interface (SNMP object ifInErrors.1) and the Serial 0/0 interface (SNMP object ifInErrors.2). In the first case, the RMON event looks for a delta (relative) rise in interface errors in a 60-second period, and a falling threshold of 5 errors per 60 seconds. In the second case,

170

Chapter 5: IP Services

the numbers are different and the thresholds are absolute, but the idea is the same. In each case, the RMON events drive RMON alarms 1, 2, 3, or 4, depending on which threshold is crossed. Example 5-10

RMON Configuration Example

rmon event 1 log trap public description Fa0.0RisingErrors owner config rmon event 2 log trap public description Fa0.0FallingErrors owner config rmon event 3 log trap public description Se0.0RisingErrors owner config rmon event 4 log trap public description Se0.0FallingErrors owner config rmon alarm 11 ifInErrors.1 60 delta rising-threshold 10 1 falling-threshold 5 2 owner config rmon alarm 20 ifInErrors.2 60 absolute rising-threshold 20 3 falling-threshold 10 4 owner config

To monitor RMON activity and to see the configured alarms and events, use the show rmon alarm and show rmon event commands. Here’s an example of the console events that take place when the events configured above trigger the corresponding alarms: Jun of Jun of

9 12:54:14.787: %RMON-5-FALLINGTRAP: Falling trap is ifInErrors.1 has fallen below the falling-threshold 9 12:55:40.732: %RMON-5-FALLINGTRAP: Falling trap is ifInErrors.2 has fallen below the falling-threshold

generated because the value value 5 generated because the value value 10

Implementing and Using FTP on a Router You can use the Cisco IOS FTP client to send or receive files from the CLI. Cisco IOS does not support configuration as an FTP server, but you can configure a TFTP server (see the next section of this chapter for details). To transfer files using FTP from the CLI, use the command ip ftp with the appropriate options. You can specify the username and password to use for an FTP transfer using the ip ftp username and ip ftp password commands. You can also specify the source interface used for FTP transfers using the ip ftp source-interface command. To initiate an FTP transfer, use the copy command with the ftp keyword in either the source or destination argument. For example, to send the startup configuration file on a router to an FTP server at 10.10.200.1, where it will be stored as r8-startup-config, the transaction is shown in Example 5-11. Example 5-11

Using FTP to Copy a Configuration File

R8# copy startup-config ftp: Address or name of remote host []? 10.10.200.1 Destination filename [r8-confg]? r8-startup-config Writing r8-startup-config

!

3525 bytes copied in 0.732 secs

Web Cache Communication Protocol

171

FTP can also be used to send an exception dump to an FTP server in the event of a crash. Example 5-12 shows a router configured to send an exception dump of 65536 bytes to 172.30.19.63 using the username JoeAdmin and password c1sco: Example 5-12

Using FTP to Send an Exception Dump

ip ftp username JoeAdmin ip ftp password c1sco ! exception protocol ftp exception region-size 65536 exception dump 172.30.19.63

Lastly, you can set the router for passive-mode FTP connections by configuring the ip ftp passive command.

Implementing a TFTP Server on a Router TFTP is commonly used for IOS and configuration file transfers on routers and switches. Cisco IOS supports configuring a TFTP server on a router, and the process is straightforward. To enable TFTP, issue the tftp-server command, which has several arguments. You can specify the memory region where the file resides (typically flash, but other regions are supported), the filename, and an access list for controlling which hosts can access the file. Here’s an example that shows the commands to permit TFTP access to flash:c1700-advipservicesk9-mz.124-23.bin to hosts that are identified by access list 11. This example also shows how the alias command-line option can be used to make the file available with a name other than the one that it has natively in flash, specifically supersecretfile.bin: tftp-server flash:c1700-advipservicesk9-mz.124-23.bin alias supersecretfile.bin 11

Implementing Secure Copy Protocol Secure Copy Protocol (SCP) is a service you can enable on a Cisco IOS router or switch to provide file copy services. SCP uses SSL (TCP port 443) for its transport protocol. It enables file transfer using the IOS copy command. SCP requires AAA for user authentication and authorization. Therefore, you must enable AAA before turning on SCP. In particular, because copy is an exec command, you must configure the aaa authorization command with the exec option. After you’ve enabled AAA, use the command ip scp server enable to turn on the SCP server.

172

Chapter 5: IP Services

Implementing HTTP and HTTPS Access Cisco IOS routers and switches support web access for administration, through both HTTP and HTTPS. Enabling HTTP access requires the ip http server global configuration command. HTTP access defaults to TCP port 80. You can change the port used for HTTP by configuring the ip http port command. You can restrict HTTP access to a router using the ip http access-class command, which applies an extended access list to connection requests. You can also specify a unique username and password for HTTP access using the ip http client username and ip http client password commands. If you choose, you can also configure HTTP access to use a variety of other access-control methods, including AAA, using ip http authentication [aaa | local | enable | tacacs] You can also configure a Cisco IOS router or switch for Secure Sockets Layer (SSL) access. By default, HTTPS uses TCP port 443, and the port is configurable in much the same way as it is with HTTP access. Enabling HTTPS access requires the http secure-server command. When you configure HTTPS access in most 12.4 IOS versions, the router or switch automatically disable HTTP access, if it has been configured. However, you should disable it manually if the router does not do it for you. HTTPS router access also gives you the option of specifying the cipher suite of your choice. This is the combination of encryption methods that the router will enable for HTTPS access. By default, all methods are enabled, as shown in this sample show command output. Example 5-13

HTTPS Configuration Output on a Router

R8# sh ip http server secure status HTTP secure server status: Enabled HTTP secure server port: 443 HTTP secure server ciphersuite: 3des-ede-cbc-sha des-cbc-sha rc4-128-md5 rc4-128-sha HTTP secure server client authentication: Disabled HTTP secure server trustpoint: HTTP secure server active session modules: ALL R8#

Implementing Telnet Access Telnet is such a ubiquitous method of access on Cisco IOS routers and switches that it needs little coverage here. Still, a few basic points are in order. Telnet requires a few configuration specifics to work. On the vty lines, the login command (or a variation of it such as login local) must be configured. If a login command is not configured, the router or switch will refuse all Telnet connection attempts.

Web Cache Communication Protocol

173

By default, Telnet uses TCP port 23. However, you can configure the vty lines to use rotary groups, also known as rotaries, to open access on other ports. If you configure this option, you should use an extended access list to enforce connection on the desired ports. By default, rotaries support connections on a number of ports. For example, if you configure rotary 33 on the vty lines, the router will accept Telnet connections on ports 3033, 5033, and 7033. Therefore, filtering undesired ports is prudent. Remember that applying access lists to vty lines requires the access-class list in command.

Implementing SSH Access Secure Shell (SSH) is much more secure than Telnet because it uses SSL rather than clear text. Therefore, today, nearly all Cisco router and switch deployments use SSH rather than Telnet for secure access. Enabling SSH on a Cisco router is a four-step process. This is because SSH requires a couple of items to be configured before you can enable SSH itself, and those prerequisites are not intuitive. The steps in configuring SSH are as follows: Step 1

Configure a hostname using the hostname command.

Step 2

Configure a domain name using the ip domain-name command.

Step 3

Configure RSA keys using the crypto key generate rsa command.

Step 4

Configure the terminal lines to permit SSH access using the transport input ssh command.

SSH supports rotaries on vty lines just as Telnet does, so you can use rotaries to specify the port or ports on which SSH access is permitted on vty lines.

174

Chapter 5: IP Services

Foundation Summary This section lists additional details and facts to round out the coverage of the topics in this chapter. Unlike most of the Cisco Press Exam Certification Guides, this “Foundation Summary” does not repeat information presented in the “Foundation Topics” section of the chapter. Please take the time to read and study the details in the “Foundation Topics” section of the chapter, as well as review items noted with a Key Topic icon. Table 5-5 lists the protocols mentioned in this chapter and their respective standards documents. Table 5-5

Protocols and Standards for Chapter 5 Name

Standardized In

ARP

RFC 826

Proxy ARP

RFC 1027

RARP

RFC 903

BOOTP

RFC 951

DHCP

RFC 2131

DHCP FQDN option

Internet-Draft

HSRP

Cisco proprietary

VRRP

RFC 3768

GLBP

Cisco proprietary

CDP

Cisco proprietary

NTP

RFC 1305

Syslog

RFC 5424

SNMP Version 1

RFCs 1155, 1156, 1157, 1212, 1213, 1215

SNMP Version 2

RFCs 1902–1907, 3416

SNMP Version 2c

RFC 1901

SNMP Version 3

RFCs 2578–2580, 3410–3415

Good Starting Point:

RFC 3410

Table 5-6 lists some of the most popular Cisco IOS commands related to the topics in this chapter.

Foundation Summary

Table 5-6

175

Command Reference for Chapter 5 Command

Description

ip dhcp pool name

Creates DHCP pool

default-router address [address2...address8]

DHCP pool subcommand to list the gateways

dns-server address [address2...address8]

DHCP pool subcommand to list DNS servers

lease {days [hours][minutes] | infinite}

DHCP pool subcommand to define the lease length

network network-number [mask | prefix-length]

DHCP pool subcommand to define IP addresses that can be assigned

ip dhcp excluded-address [low-address highaddress]

DHCP pool subcommand to disallow these addresses from being assigned

host address [mask | prefix-length]

DHCP pool subcommand, used with hardware-address or client-identifier, to predefine a single host’s IP address

hardware-address hardware-address type

DHCP pool subcommand to define MAC address; works with host command

show ip dhcp binding [ip-address]

Lists addresses allocated by DHCP

show ip dhcp server statistics

Lists stats for DHCP server operations

standby [group-number] ip [ip-address [secondary]]

Interface subcommand to enable an HSRP group and define the virtual IP address

track object-number interface type-number {line-protocol | ip routing}

Configures a tracking object that can be used by HSRP, VRRP, or GLBP to track the status of an interface

standby [group-number] preempt [delay {minimum delay | reload delay | sync delay}]

Interface subcommand to enable pre-emption and set delay timers

show track [object-number [brief] | interface [brief] | ip route [brief] | resolution | timers]

Displays status of tracked objects

standby [group-number] priority priority

Interface subcommand to set the HSRP group priority for this router

standby [group-number] timers [msec] hellotime [msec] holdtime

Interface subcommand to set HSRP group timers

standby [group-number] track object-number

Interface subcommand to enable HSRP to track defined objects, usually for the purpose of switching active routers on an event related to that object

show standby [type number [group]] [brief | all]

Lists HSRP statistics

ntp peer ip-address [version number] [key keyid] [source interface] [prefer]

Global command to enable symmetric active mode NTP continues

176

Chapter 5: IP Services

Table 5-6

Command Reference for Chapter 5 (Continued) Command

Description

ntp server ip-address [version number] [key keyid] [source interface] [prefer]

Global command to enable static client mode NTP

ntp broadcast [version number]

Interface subcommand on an NTP server to cause NTP broadcasts on the interface

ntp broadcast client

Interface subcommand on an NTP client to cause it to listen for NTP broadcasts

ntp master [stratum]

Global command to enable NTP server

show ntp associations

Lists associations with other NTP servers and clients

show ntp status

Displays synchronization status, stratum level, and other basic information

logging trap level

Sets the severity level for syslog messages; arguments are 0–7, where 0=emergencies, 1=alerts, 2=critical, 3=errors, 4=warnings, 5=notifications, 6=informational, 7=debugging (default)

logging host {{ip-address | hostname} | {ipv6 ipv6-address | hostname}} [transport {udp [port port-number] | tcp [port port-number]}] [alarm [severity]]

Configures the IP or IPv6 address or hostname to which to send syslog messages and permits setting the transport protocol and port number

ip wccp {web-cache | service-number} [servicelist service-access-list] [mode {open | closed}] [group-address multicast-address] [redirect-list access-list] [group-list access-list] [password [07] password]

Enables WCCP and configures filtering and service parameters

ip wccp {web-cache | service-number} redirect {in | out}

Interface configuration command to enable WCCP and configure it for outbound or inbound service

show ip wccp

Displays WCCP configuration settings and statistics

snmp-server enable traps

Enables sending of all types of traps available on the router or switch.

snmp-server host {hostname | ip-address} [vrf vrf-name] [traps | informs] [version {1 | 2c | 3 [auth | noauth | priv]}] community-string [udpport port] [notification-type]

Configures the SNMP server to send traps or informs to a particular host, along with options for setting the SNMP version for traps and the UDP port (default is 162). The notification-type field specifies the types of traps to send; if no types are specified, all available categories of traps will be sent.

Foundation Summary

Table 5-6

177

Command Reference for Chapter 5 (Continued) Command

Description

snmp-server community string [view viewname] [ro | rw] [access-list-number]

Sets the read-only or read-write community string and access list for host filtering for access to SNMP reads and writes on the router or switch.

show snmp mib ifmib ifindex interface-id

Shows the router’s interface ID for a particular interface. Particularly useful for RMON configuration.

ip sla monitor operation-index

Enters IP SLA monitor configuration mode for an individual monitor function.

type [jitter | udp-echo | echo protocol icmpecho | dns | ftp operation | http operation | mpls ping ipv4 | pathecho | pathjitter | tcpconnect | voip delay post-dial | udp-jitter | udp-jitter codec]

Configures the IP SLA monitor type with options (not shown) including source and destination IP address and source and destination port number, plus other relevant options to the particular type.

ip sla key-chain key-chain-name

Configures a keychain for MD5 authentication of IP SLA operations.

ip sla monitor schedule operation-number [life {forever | seconds}] [start-time {hh:mm[:ss] [month day | day month] | pending | now | after hh:mm:ss}] [ageout seconds] [recurring]

Configures the schedule for a particular IP SLA monitor. If the IP SLA monitor is deleted from the configuration, the schedule is also deleted.

ip sla monitor responder

Enables the IP SLA responder function globally. More specific options for this command may be configured for specific responder types, ports, and so on.

show ip sla monitor statistics [operation] detail

Shows the statistics for a specified IP SLA operation or all configured IP SLA operations.

show ip sla responder

Shows currently configured IP SLA responders and recent activity (source IP address, and so forth).

ip ssh [timeout seconds | authentication-retries integer]

Enables SSH access.

crypto key generate rsa

Generates RSA keys. Required for SSH configuration.

transport input ssh

In vty configuration mode, permits SSH connections.

ip http server

Enables HTTP server.

ip http secure-server

Enables HTTPS server. continues

178

Chapter 5: IP Services

Table 5-6

Command Reference for Chapter 5 (Continued) Command

Description

ip traffic-export profile profile-name

Enables and enters configuration mode for a RITE profile.

ip traffic-export apply profile-name

Applies a RITE profile to an interface.

event manager applet applet-name [class classoptions] [trap]

Enters EEM applet configuration mode.

event cli pattern regular-expression {[default] [enter] [questionmark] [tab]} [sync {yes | no skip {yes | no}] [mode variable] [occurs numoccurrences] [period period-value] [maxrun maxruntime-number]

Configures EEM to match a CLI command string.

ip flow-top-talkers

NetFlow aggregator. Aggregates traffic for unclassified top talkers.

flow monitor flow-name

Enters configuration mode for a NetFlow monitor.

flow exporter exporter-name

Configures a NetFlow exporter and the destination server to which to send NetFlow information for a particular flow monitor.

rmon event

Configures an RMON event to monitor a particular SNMP object, along with rising and falling thresholds.

rmon alarm

Configures an alarm action for an RMON event’s rising or falling threshold.

copy

With FTP option in the source or destination field, copies a file to or from an FTP server.

tftp-server flash [partition-number:] filename1 [alias filename2] [access-list-number]

Configures a TFTP server on the router to serve a file, optionally with an alias, and optionally through an ACL.

aaa new-model

Enables AAA on the router.

aaa authentication

Configures AAA authentication methods.

aaa authorization

Configures AAA authorization methods.

ip scp server enable

Enables the SCP server on the router. Requires AAA authentication and AAA authorization to be configured.

Memory Builders

179

Memory Builders The CCIE Routing and Switching written exam, like all Cisco CCIE written exams, covers a fairly broad set of topics. This section provides some basic tools to help you exercise your memory about some of the broader topics covered in this chapter.

Fill In Key Tables from Memory Appendix G, “Key Tables for CCIE Study,” on the CD in the back of this book contains empty sets of some of the key summary tables in each chapter. Print Appendix G, refer to this chapter’s tables in it, and fill in the tables from memory. Refer to Appendix H, “Solutions for Key Tables for CCIE Study,” on the CD to check your answers.

Definitions Next, take a few moments to write down the definitions for the following terms: HSRP, VRRP, GLBP, ARP, RARP, proxy ARP, BOOTP, DHCP, NTP symmetric active mode, NTP server mode, NTP client mode, NTP, virtual IP address, VRRP Master router, SNMP agent, SNMP manager, Get, GetNext, GetBulk, MIB-I, MIB-II, Response, Trap, Set, Inform, SMI, MIB, MIB walk, lead content engine Refer to the glossary to check your answers.

Further Reading More information about several of the topics in this chapter can be easily found in a large number of books and online documentation. The RFCs listed in Table 5-5 of the “Foundation Summary” section also provide a great deal of background information for this chapter. Here are a few references for more information about some of the less popular topics covered in this chapter: ■

Proxy ARP—http://www.cisco.com/en/US/tech/tk648/tk361/ technologies_tech_note09186a0080094adb.shtml.

■

GLBP—http://www.cisco.com/en/US/docs/ios/12_2t/12_2t15/feature/guide/ft_glbp.html.

■

VRRP—http://www.cisco.com/en/US/docs/ios/12_0st/12_0st18/feature/guide/ st_vrrpx.html.

■

SNMP—Any further reading of SNMP-related RFCs should begin with RFC 3410, which provides a great overview of the releases and points to the more important of the vast number of SNMP-related RFCs.

Blueprint topics covered in this chapter: This chapter covers the following subtopics from the Cisco CCIE Routing and Switching written exam blueprint. Refer to the full blueprint in Table I-1 in the Introduction for more details on the topics covered in each chapter and their context within the blueprint. ■

Classful and Classless Routing Protocols

■

Concepts of Policy Routing

■

Performance Routing (PfR) and Optimized Edge Routing (OER)

■

IPv4 Tunneling and GRE

CHAPTER

6

IP Forwarding (Routing) Chapter 6 begins the largest part of the book. This part of the book, containing Chapters 7 through 11, focuses on the topics that are the most important and popular for both the CCIE Routing and Switching written and practical (lab) exams. Chapter 6 begins with coverage of the details of the forwarding plane—the actual forwarding of IP packets. This process of forwarding IP packets is often called IP routing, or simply routing. Also, many people also refer to IP routing as the data plane, meaning the plane (topic) related to the end-user data. Chapters 7 through 11 cover the details of the IP control plane. In contrast to the term data plane, the control plane relates to the communication of control information—in short, routing protocols like OSPF and BGP. These chapters cover the routing protocols on the exam, one chapter per routing protocol, plus an additional chapter on redistribution and route summarization.

“Do I Know This Already?” Quiz Table 6-1 outlines the major headings in this chapter and the corresponding “Do I Know This Already?” quiz questions. Table 6-1

“Do I Know This Already?” Foundation Topics Section-to-Question Mapping Foundation Topics Section

Questions Covered in This Section

IP Forwarding

1–6

Multilayer Switching

7–8

Policy Routing

9–10

Optimized Edge Routing and Performance Routing

11

Total Score

Score

182

Chapter 6: IP Forwarding (Routing)

To best use this pre-chapter assessment, remember to score yourself strictly. You can find the answers in Appendix A, “Answers to the ‘Do I Know This Already?’ Quizzes.” 1.

2.

3.

4.

What command is used to enable CEF globally for IP packets? a.

enable cef

b.

ip enable cef

c.

ip cef

d.

cef enable

e.

cef enable ip

f.

cef ip

Which of the follow triggers an update to a CEF FIB? a.

Receipt of a Frame Relay InARP message with previously unknown information

b.

Receipt of a LAN ARP reply message with previously unknown information

c.

Addition of a new route to the IP routing table by EIGRP

d.

Addition of a new route to the IP routing table by adding an ip route command

e.

The removal of a route from the IP routing table by EIGRP

Router1 has a Frame Relay access link attached to its s0/0 interface. Router1 has a PVC connecting it to Router3. What action triggers Router3 to send an InARP message over the PVC to Router1? a.

Receipt of a CDP multicast on the PVC connected to Router1

b.

Receipt of an InARP request from Router1

c.

Receipt of a packet that needs to be routed to Router1

d.

Receipt of a Frame Relay message stating the PVC to Router1 is up

Three routers are attached to the same Frame Relay network, have a full mesh of PVCs, and use IP addresses 10.1.1.1/24 (R1), 10.1.1.2/24 (R2), and 10.1.1.3 (R3). R1 has its IP address configured on its physical interface; R2 and R3 have their IP addresses configured on multipoint subinterfaces. Assuming all the Frame Relay PVCs are up and working, and the router interfaces have been administratively enabled, which of the following is true? a.

R1 can ping 10.1.1.2.

b.

R2 cannot ping 10.1.1.3.

c.

R3 can ping 10.1.1.2.

d.

In this case, R1 must rely on mapping via InARP to be able to ping 10.1.1.3.

“Do I Know This Already?” Quiz

5.

6.

7.

183

Three routers are attached to the same Frame Relay network, with a partial mesh of PVCs: R1-R2 and R1-R3. The routers use IP addresses 10.1.1.1/24 (R1), 10.1.1.2/24 (R2), and 10.1.1.3/24 (R3). R1 has its IP address configured on its physical interface; R2 has its IP address configured on a multipoint subinterface; and R3 has its IP address configured on a point-topoint subinterface. Assuming all the Frame Relay PVCs are up and working, and the router interfaces have been administratively enabled, which of the following is true? Assume no frame-relay map commands have been configured. a.

R1 can ping 10.1.1.2.

b.

R2 can ping 10.1.1.3.

c.

R3 can ping 10.1.1.1.

d.

R3’s ping 10.1.1.2 command results in R3 not sending the ICMP Echo packet.

e.

R2’s ping 10.1.1.3 command results in R2 not sending the ICMP Echo packet.

Router1 has an OSPF-learned route to 10.1.1.0/24 as its only route to a subnet on class A network 10.0.0.0. It also has a default route. When Router1 receives a packet destined for 10.1.2.3, it discards the packet. Which of the following commands would make Router1 use the default route for those packets in the future? a.

ip classless subcommand of router ospf

b.

no ip classful subcommand of router ospf

c.

ip classless global command

d.

no ip classless global command

e.

no ip classful global command

Which of the following commands is used on a Cisco IOS Layer 3 switch to use the interface as a routed interface instead of a switched interface? a.

ip routing global command

b.

ip routing interface subcommand

c.

ip address interface subcommand

d.

switchport access layer-3 interface subcommand

e.

no switchport interface subcommand

184

Chapter 6: IP Forwarding (Routing)

8.

9.

10.

On a Cisco 3550 switch with Enterprise Edition software, the first line of the output of a show interface vlan 55 command lists the state as “Vlan 55 is down, line protocol is down down.” Which of the following might be causing that state to occur? a.

VLAN interface has not been no shut yet.

b.

The ip routing global command is missing from the configuration.

c.

On at least one interface in the VLAN, a cable that was previously plugged in has been unplugged.

d.

VTP mode is set to transparent.

e.

The VLAN has not yet been created on this switch.

Imagine a route map used for policy routing, in which the route map has a set interface default serial0/0 command. Serial0/0 is a point-to-point link to another router. A packet arrives at this router, and the packet matches the policy routing route-map clause whose only set command is the one just mentioned. Which of the following general characterizations is true? a.

The packet will be routed out interface s0/0; if s0/0 is down, it will be routed using the default route from the routing table.

b.

The packet will be routed using the default route in the routing table; if there is no default, the packet will be routed out s0/0.

c.

The packet will be routed using the best match of the destination address with the routing table; if no match is made, the packet will be routed out s0/0.

d.

The packet will be routed out interface s0/0; if s0/0 is down, the packet will be discarded.

Router1 has an fa0/0 interface and two point-to-point WAN links back to the core of the network (s0/0 and s0/1, respectively). Router1 accepts routing information only over s0/0, which Router1 uses as its primary link. When s0/0 fails, Router1 uses policy routing to forward the traffic out the relatively slower s0/1 link. Which of the following set commands in Router1’s policy routing route map could have been used to achieve this function? a.

set ip default next-hop

b.

set ip next-hop

c.

set ip default interface

d.

set ip interface

“Do I Know This Already?” Quiz

11.

Which of the following are conditions or attributes of PfR? a.

Requires CEF to be enabled

b.

Doesn’t support MPLS circuits

c.

Operates automatically in Cisco IOS 12.4(9)T and later, with no configuration

d.

Supports passive and active monitoring

e.

Uses IP SLA and NetFlow to gather performance information

185

186

Chapter 6: IP Forwarding (Routing)

Foundation Topics

IP Forwarding IP forwarding, or IP routing, is simply the process of receiving an IP packet, making a decision of where to send the packet next, and then forwarding the packet. The forwarding process needs to be relatively simple, or at least streamlined, for a router to forward large volumes of packets. Ignoring the details of several Cisco optimizations to the forwarding process for a moment, the internal forwarding logic in a router works basically as shown in Figure 6-1. Forwarding Process at Router3, Destination Server1

Figure 6-1

172.31.0.0, /24 Prefixes Client 3 Telnet Server – 172.31.11.201

103.3 E0/0

GW is

S0/0.3333

13.1

R3

R1

103.3

IP Routing Table

Solid lines show packet flow

3

Compare packet destination with routing table

172.31.103.0/24 Connected 172.31.13.0/24 Connected 172.31.11.0/24 172.31.13.1

E0/0 S0/0.3333 S0/0.3333

4 5 IP Packet

Adjacency Information

Update TTL and Header Checksum

IP Packet

172.31.13.1

DLCI 100

Type = 0x0800

2

6

Eth. Eth. IP Packet Header Trailer

FR FR IP Packet Header Trailer

1

Insert into Header

Passed FCS

Eth. Eth. IP Packet Header Trailer

The following list summarizes the key steps shown in Figure 6-1. 1.

A router receives the frame and checks the received frame check sequence (FCS); if errors occurred, the frame is discarded. The router makes no attempt to recover the lost packet.

2.

If no errors occurred, the router checks the Ethernet Type field for the packet type, and extracts the packet. The Data Link header and trailer can now be discarded.

3.

Assuming an IP packet, the router checks its IP routing table for the most specific prefix match of the packet’s destination IP address.

IP Forwarding

187

4.

The matched routing table entry includes the outgoing interface and next-hop router; this information points the router to the adjacency information needed to build a new Data Link frame.

5.

Before creating a new frame, the router updates the IP header TTL field, requiring a recomputation of the IP header checksum.

6.

The router encapsulates the IP packet in a new Data Link header (including the destination address) and trailer (including a new FCS) to create a new frame.

The preceding list is a generic view of the process; next, a few words on how Cisco routers can optimize the routing process by using Cisco Express Forwarding (CEF).

Process Switching, Fast Switching, and Cisco Express Forwarding Steps 3 and 4 from the generic routing logic shown in the preceding section are the most computation-intensive tasks in the routing process. A router must find the best route to use for every packet, requiring some form of table lookup of routing information. Also, a new Data Link header and trailer must be created, and the information to put in the header (like the destination Data Link address) must be found in another table. Cisco has created several different methods to optimize the forwarding processing inside routers, termed switching paths. This section examines the two most likely methods to exist in Cisco router networks today: fast switching and CEF. With fast switching, the first packet to a specific destination IP address is process switched, meaning that it follows the same general algorithm as in Figure 6-1. With the first packet, the router adds an entry to the fast-switching cache, sometimes called the route cache. The cache has the destination IP address, the next-hop information, and the data link header information that needs to be added to the packet before forwarding (as in Step 6 in Figure 6-1). Future packets to the same destination address match the cache entry, so it takes the router less time to process and forward the packet. Although it is much better than process switching, fast switching has a few drawbacks. The first packet must be process switched. The cache entries are timed out relatively quickly, because otherwise the cache could get overly large as it has an entry per each destination address, not per destination subnet/prefix. Also, load balancing can only occur per destination with fast switching. CEF overcomes the main shortcoming of fast switching. CEF optimizes the route lookup process by using a construct called the Forwarding Information Base (FIB). The FIB contains information about all the known routes in the routing table. Rather than use a table that is updated when new flows appear, as did the Cisco earlier fast-switching technology, CEF loads FIB entries as routes are added and removed from the routing table. CEF does not have to time out the entries in the FIB, does not require the first packet to a destination to be process switched, and allows much more effective load balancing over equal-cost routes.

188

Chapter 6: IP Forwarding (Routing)

When a new packet arrives, CEF routers first search the FIB. Cisco designed the CEF FIB structure as a special kind of tree, called an mtrie, that significantly reduces the time taken to match the packet destination address to the right CEF FIB entry. The matched FIB entry points to an entry in the CEF adjacency table. The adjacency table lists the outgoing interface, along with all the information needed to build the Data Link header and trailer before sending the packet. When a router forwards a packet using CEF, it easily and quickly finds the corresponding CEF FIB entry, after which it has a pointer to the adjacency table entry— which tells the router how to forward the packet. Table 6-2 summarizes a few key points about the three main options for router switching paths. Table 6-2

Matching Logic and Load-Balancing Options for Each Switching Path Switching Path

Tables that Hold the Forwarding Information

Load-Balancing Method

Process switching

Routing table

Per packet

Fast switching

Fast-switching cache (per flow route cache)

Per destination IP address

CEF

FIB and adjacency tables

Per a hash of the packet source and destination, or per packet

The ip cef global configuration command enables CEF for all interfaces on a Cisco router. The no ip route-cache cef interface subcommand can then be used to selectively disable CEF on an interface. On many of the higher-end Cisco platforms, CEF processing can be distributed to the linecards. Similarly, Cisco multilayer switches use CEF for Layer 3 forwarding, by loading CEF tables into the forwarding ASICs.

Building Adjacency Information: ARP and Inverse ARP The CEF adjacency table entries list an outgoing interface and a Layer 2 and Layer 3 address reachable via that interface. The table also includes the entire data link header that should be used to reach that next-hop (adjacent) device. The CEF adjacency table must be built based on the IP routing table, plus other sources. The IP routing table entries include the outgoing interfaces to use and the next-hop device’s IP address. To complete the adjacency table entry for that next hop, the router needs to know the Data Link layer address to use to reach the next device. Once known, the router can build the CEF adjacency table entry for that next-hop router. For instance, for Router3 in Figure 6-1 to reach next-hop router 172.31.13.1 (Router1), out interface s0/0.3333, Router3 needed to know the right Data-Link connection identifier (DLCI) to use. So, to build the adjacency table entries, CEF uses the IP ARP cache, Frame Relay mapping information, and other sources of Layer 3-to-Layer 2 mapping information.

IP Forwarding

189

First, a quick review of IP ARP. The ARP protocol dynamically learns the MAC address of another IP host on the same LAN. The host that needs to learn the other host’s MAC address sends an ARP request, sent to the LAN broadcast address, hoping to receive an ARP reply (a LAN unicast) from the other host. The reply, of course, supplies the needed MAC address information. Frame Relay Inverse ARP IP ARP is widely understood and relatively simple. As a result, it is hard to ask difficult questions about ARP on an exam. However, the topics of Frame Relay Inverse ARP (InARP), its use, defaults, and when static mapping must be used lend themselves to being the source of tricky exam questions. So, this section covers Frame Relay InARP to show some of the nuances of when and how it is used. InARP discovers the DLCI to use to reach a particular adjacent IP address. However, as the term Inverse ARP implies, the process differs from ARP on LANs; with InARP, routers already know the Data Link address (DLCI), and need to learn the corresponding IP address. Figure 6-2 shows an example InARP flow. Figure 6-2

Frame Relay InARP 172.31.124.4

172.31.124.2

R4

R2 LMI Status: DLCI 200 up

LMI Status: DLCI 400 up

InARP: I am 172.31.124.4

InARP: I am 172.31.124.2

Unlike on LANs, a packet does not need to arrive at the router to trigger the InARP protocol; instead, an LMI status message triggers InARP. After receiving an LMI PVC Up message, each router announces its own IP address over the VC, using an InARP message, as defined in RFC 1293. Interestingly, if you disable LMI, then the InARP process no longer works, because nothing triggers a router to send an InARP message. NOTE In production Frame Relay networks, the configuration details are chosen to purposefully avoid some of the pitfalls that are covered in the next several pages of this chapter. For example, when using mainly point-to-point subinterfaces, with a different subnet per VC, all the problems described in the rest of the Frame Relay coverage in this chapter can be avoided.

While InARP itself is relatively simple, implementation details differ based on the type of subinterface used in a router. For a closer look at implementation, Figure 6-3 shows an example Frame Relay topology with a partial mesh and a single subnet, in which each router uses a different interface type for its Frame Relay configuration. (You would not typically choose to configure

190

Chapter 6: IP Forwarding (Routing)

Frame Relay on physical, point-to-point and multipoint subinterfaces in the same design—indeed, it wreaks havoc with routing protocols if you do so. This example does so just to show in more detail in the examples how InARP really works.) Example 6-1 points out some of the basic show and debug commands related to Frame Relay InARP, and one of the oddities about InARP relating to point-to-point subinterfaces. Figure 6-3

Frame Relay Topology for Frame Relay InARP Examples 172.31.134.0/24 Pt-Pt 134.3 Mpt 134.1

R3 DLCI 300

R1 DLCI 100

Physical 134.4

R4

DLCI 400

NOTE All figures with Frame Relay networks in this book use Global DLCI conventions unless otherwise stated. For instance, in Figure 6-3, DLCI 300 listed beside Router3 means that, due to Local DLCI assignment conventions by the service provider, all other routers (like Router4) use DLCI 300 to address their respective VCs back to Router3.

Example 6-1

Frame Relay InARP show and debug Commands

! First, Router1 configures Frame Relay on a multipoint subinterface. Router1# sh run ! Lines omitted for brevity interface Seria l0/0 e n c a ps u l a t i o n f r a m e - r e l a y interface Serial0/0.11 multipoint i p a d d r e s s 1 7 2 . 3 1 . 1 3 4 .1 2 5 5 . 2 5 5 . 2 5 5 . 0 frame-relay interface-dlci 300 f r a m e - r e l a y i n t e r f a c e - d l c i 40 0 ! Lines omitted for brevity ! Next, the serial interface is shut and no shut, and the earlier InARP entries ! are cleared, so the example can show the InARP process. Router1# conf t Enter configuration commands, one per line.

End with CNTL/Z.

Router1(config)# int s 0/0 Router1(config-if)# do clear frame-relay inarp Router1(config-if)# shut Router1(config-if)# no shut Router1(config-if)# ^Z

IP Forwarding

Example 6-1

191

Frame Relay InARP show and debug Commands (Continued)

! Messages resulting from the debug frame-relay event command show the ! received InARP messages on Router1. Note the hex values 0xAC1F8603 and ! 0xAC1F8604, which in decimal are 172.31.134.3 and 172.31.134.4 (Router3 ! and Router4, respectively). Router1# debug frame-relay events *Mar

1 00:09:45.334: Serial0/0.11: FR ARP input

*Mar

1 00:09:45.334: datagramstart = 0x392BA0E, datagramsize = 34

*Mar

1 00:09:45.334: FR encap = 0x48C10300

*Mar

1 00:09:45.334: 80 00 00 00 08 06 00 0F 08 00 02 04 00 09 00 00

*Mar

1 00:09:45.334: AC 1F 86 03 48 C1 AC 1F 86 01 01 02 00 00

*Mar

1 00:09:45.334:

*Mar

1 00:09:45.334: Serial0/0.11: FR ARP input

*Mar

1 00:09:45.334: datagramstart = 0x392B8CE, datagramsize = 34

*Mar

1 00:09:45.338: FR encap = 0x64010300

*Mar

1 00:09:45.338: 80 00 00 00 08 06 00 0F 08 00 02 04 00 09 00 00

*Mar

1 00:09:45.338: AC 1F 86 04 64 01 AC 1F 86 01 01 02 00 00

! Next, note the show frame-relay map command output does include a “dynamic” ! keyword, meaning that the entries were learned with InARP. Router1# show frame-relay map Serial0/0.11 (up): ip 172.31.134.3 dlci 300(0x12C,0x48C0), dynamic, broadcast,, status defined, active Serial0/0.11 (up): ip 172.31.134.4 dlci 400(0x190,0x6400), dynamic, broadcast,, status defined, active ! On Router3, show frame-relay map only lists a single entry as well, but ! the format is different. Because Router3 uses a point-to-point subinterface, ! the entry was not learned with InARP, and the command output does not include ! the word “dynamic.” Also note the absence of any Layer 3 addresses. Router3# show frame-relay map Serial0/0.3333 (up): point-to-point dlci, dlci 100(0x64,0x1840), broadcast status defined, active

NOTE Example 6-1 included the use of the do command inside configuration mode. The do command, followed by any exec command, can be used from inside configuration mode to issue an exec command, without having to leave configuration mode. The example show commands from Router1 detail the fact that InARP was used; however, the last show command in Example 6-1 details how Router3 actually did not use the received InARP information. Cisco IOS Software knows that only one VC is associated with a point-to-point subinterface; any other IP hosts in the same subnet as a point-to-point subinterface can be reached only by that single DLCI. So, any received InARP information related to that DLCI is unnecessary. For instance, whenever Router3 needs to forward a packet to Router1 (172.31.134.1), or any other host in subnet 172.31.134.0/24, Router3 already knows from its configuration to send the packet

192

Chapter 6: IP Forwarding (Routing)

over the only possible DLCI on that point-to-point subinterface—namely, DLCI 100. So, although all three types of interfaces used for Frame Relay configuration support InARP by default, pointto-point subinterfaces ignore the received InARP information. Static Configuration of Frame Relay Mapping Information In Figure 6-3, Router3 already knows how to forward frames to Router4, but the reverse is not true. Router3 uses logic like this: “For packets needing to get to a next-hop router that is in subnet 172.31.124.0/24, send them out the one DLCI on that point-to-point subinterface—DLCI 100.” The packet then goes over that VC to Router1, which in turn routes the packet to Router4. In the admittedly poor design shown in Figure 6-3, however, Router4 cannot use the same kind of logic as Router3, as its Frame Relay details are configured on its physical interface. To reach Router3, Router4 needs to send frames over DLCI 100 back to Router1, and let Router1 forward the packet on to Router3. In this case, InARP does not help, because InARP messages only flow across a VC, and are not forwarded; note that there is no VC between Router4 and Router3. The solution is to add a frame-relay map command to Router4’s configuration, as shown in Example 6-2. The example begins before Router4 has added the frame-relay map command, and then shows the results after having added the command. Example 6-2

Using the frame-relay map Command—Router4

! Router4 only lists a single entry in the show frame-relay map command ! output, because Router4 only has a single VC, which connects back to Router1. ! With only 1 VC, Router4 could only have learned of 1 other router via InARP. Router4# sh run ! lines omitted for brevity interface Serial0/0 ip address 172.31.134.4 255.255.255.0 encapsulation frame-relay Router4# show frame-relay map Serial0/0 (up): ip 172.31.134.1 dlci 100(0x64,0x1840), dynamic, broadcast,, status defined, active ! Next, proof that Router4 cannot send packets to Router3’s Frame Relay IP address. Router4# ping 172.31.134.3 Type escape sequence to abort. Sending 5, 100-byte ICMP Echos to 172.31.134.3, timeout is 2 seconds: ..... Success rate is 0 percent (0/5) ! Next, static mapping information is added to Router4 using the frame-relay map ! interface subcommand. Note that the command uses DLCI 100, so that any packets ! sent by Router4 to 172.31.134.3 (Router3) will go over the VC to Router1, which ! will then need to route the packet to Router3. The broadcast keyword tells

IP Forwarding

Example 6-2

193

Using the frame-relay map Command—Router4 (Continued)

! Router4 to send copies of broadcasts over this VC. Router4# conf t Enter configuration commands, one per line.

End with CNTL/Z.

Router4(config)# int s0/0 Router4(config-if)# frame-relay map ip 172.31.134.3 100 broadcast Router4(config-if)# ^Z Router4# ping 172.31.134.3 Type escape sequence to abort. Sending 5, 100-byte ICMP Echos to 172.31.134.3, timeout is 2 seconds: !!!!! Success rate is 100 percent (5/5), round-trip min/avg/max = 20/20/20 ms

NOTE Remember, Router3 did not need a frame-relay map command, due to the logic used for a point-to-point subinterface. Keep in mind that you probably would not choose to build a network like the one shown in Figure 6-3 using different subinterface types on the remote routers, nor would you typically put all three nonfully meshed routers into the same subnet unless you were seriously constrained in your IP address space. In cases where you do use a topology like that shown in Figure 6-3, you can use the configuration described in the last few pages. Alternatively, if both Router3 and Router4 had used multipoint subinterfaces, they would both have needed frame-relay map commands, because these two routers could not have heard InARP messages from the other router. However, if both Router3 and Router4 had used point-to-point subinterfaces, neither would have required a frame-relay map command, due to the “use this VC to reach all addresses in this subnet” logic. Disabling InARP In most cases for production networks, using InARP makes sense. However, InARP can be disabled on multipoint and physical interfaces using the no frame-relay inverse-arp interface subcommand. InARP can be disabled for all VCs on the interface/subinterface, all VCs on the interface/subinterface for a particular Layer 3 protocol, and even for a particular Layer 3 protocol per DLCI. Interestingly, the no frame-relay inverse-arp command not only tells the router to stop sending InARP messages, but also tells the router to ignore received InARP messages. For instance, the no frame-relay inverse-arp ip 400 subinterface subcommand on Router1 in Example 6-2 not only prevents Router1 from sending InARP messages over DLCI 400 to Router4, but also causes Router1 to ignore the InARP received over DLCI 400.

194

Chapter 6: IP Forwarding (Routing)

Table 6-3 summarizes some of the key details about Frame Relay Inverse ARP settings in IOS. Facts and Behavior Related to InARP

Table 6-3

Fact/Behavior

Point-to-Point

Multipoint or Physical

Does InARP require LMI?

Always

Always

Is InARP enabled by default?

Yes

Yes

Can InARP be disabled?

No

Yes

Ignores received InARP messages?

Always1

When InARP is disabled

1Point-to-point interfaces ignore InARP messages because of their “send all packets for addresses in this subnet using

the only DLCI on the subinterface” logic.

Classless and Classful Routing So far this chapter has reviewed the basic forwarding process for IP packets in a Cisco router. The logic requires matching the packet destination with the routing table, or with the CEF FIB if CEF is enabled, or with other tables for the other options Cisco uses for route table lookup. (Those options include fast switching in routers and NetFlow switching in multilayer switches, both of which populate an optimized forwarding table based on flows, but not on the contents of the routing table.) Classless routing and classful routing relate to the logic used to match the routing table, specifically for when the default route is used. Regardless of the use of any optimized forwarding methods (for instance, CEF), the following statements are true about classless and classful routing: ■

Classless routing—When a default route exists, and no specific match is made when comparing the destination of the packet and the routing table, the default route is used.

■

Classful routing—When a default route exists, and the class A, B, or C network for the destination IP address does not exist at all in the routing table, the default route is used. If any part of that classful network exists in the routing table, but the packet does not match any of the existing subnets of that classful network, the router does not use the default route and thus discards the packet.

Typically, classful routing works well in enterprise networks only when all the enterprise routes are known by all routers, and the default is used only to reach the Internet-facing routers. Conversely, for enterprise routers that normally do not know all the routes—for instance, if a remote router has only a few connected routes to network 10.0.0.0 and a default route pointing back to a core site— classless routing is required. For instance, in an OSPF design using stubby areas, default routes are injected into the non-backbone areas, instead of advertising all routes to specific subnets. As a

Multilayer Switching

195

result, classless routing is required in routers in the stubby area, because otherwise non-backbone area routers would not be able to forward packets to all parts of the network. Classless and classful routing logic is controlled by the ip classless global configuration command. The ip classless command enables classless routing, and the no ip classless command enables classful routing. No single chapter in this book covers the details of the three uses of the terms classful and classless. Table 6-4 summarizes and compares the three uses of these terms. Table 6-4

Comparing the Use of the Terms Classless and Classful

As the Terms Pertain to . . .

Meaning of “Classless”

Addressing (Chapter 4)

Class A, B, and C rules are not used; Class A, B, and C rules are used; addresses have addresses have two parts, the prefix and host. three parts, the network, subnet, and host.

Routing (Chapter 6)

If no specific routes are matched for a given packet, the router forwards based on the default route.

The router first attempts a match of the classful network. If found, but none of the routes in that classful network matches the destination of a given packet, the default route is not used.

Routing protocols (Chapters 7-10)

Routing protocol does not need to assume details about the mask, as it is included in the routing updates; supports VLSM and discontiguous networks. Classless routing protocols: RIPv2, EIGRP, OSPF, and IS-IS.

Routing protocol does need to assume details about the mask, as it is not included in the routing updates; does not support VLSM and discontiguous networks. Classful routing protocols: RIPv1 and IGRP.

Meaning of “Classful”

Multilayer Switching Multilayer Switching (MLS) refers to the process by which a LAN switch, which operates at least at Layer 2, also uses logic and protocols from layers other than Layer 2 to forward data. The term Layer 3 switching refers specifically to the use of the Layer 3 destination address, compared to the routing table (or equivalent), to make the forwarding decision. (The latest switch hardware and software from Cisco uses CEF switching to optimize the forwarding of packets at Layer 3.)

MLS Logic Layer 3 switching configuration works similarly to router configuration—IP addresses are assigned to interfaces, and routing protocols are defined. The routing protocol configuration works just like a router; however, the interface configuration on MLS switches differs slightly from routers, using VLAN interfaces, routed interfaces, and PortChannel interfaces.

196

Chapter 6: IP Forwarding (Routing)

VLAN interfaces give a Layer 3 switch a Layer 3 interface attached to a VLAN. Cisco sometimes refers to these interfaces as switched virtual interfaces (SVIs). To route between VLANs, a switch simply needs a virtual interface attached to each VLAN, and each VLAN interface needs an IP address in the respective subnets used on those VLANs. NOTE Although it is not a requirement, the devices in a VLAN are typically configured in the same single IP subnet. However, you can use secondary IP addresses on VLAN interfaces to configure multiple subnets in one VLAN, just like on other router interfaces.

When using VLAN interfaces, the switch must take one noticeable but simple additional step when routing a packet. Like typical routers, MLS makes a routing decision to forward a packet. As with routers, the routes in an MLS routing table entry list an outgoing interface (a VLAN interface in this case), as well as a next-hop layer 3 address. The adjacency information (for example, the IP ARP table or the CEF adjacency table) lists the VLAN number and the next-hop device’s MAC address to which the packet should be forwarded—again, typical of normal router operation. At this point, a true router would know everything it needs to know to forward the packet. An MLS switch, however, then also needs to use Layer 2 logic to decide out which physical interface to physically forward the packet. The switch will simply find the next-hop device’s MAC address in the CAM and forward the frame to that address based on the CAM.

Using Routed Ports and PortChannels with MLS In some point-to-point topologies, VLAN interfaces are not required. For instance, when an MLS switch connects to a router using a cable from a switch interface to a router’s LAN interface, and the only two devices in that subnet are the router and that one physical interface on the MLS switch, the MLS switch can be configured to treat that one interface as a routed port. (Another typical topology for using router ports is when two MLS switches connect for the purpose of routing between the switches, again creating a case with only two devices in the VLAN/subnet.) A routed port on an MLS switch has the following characteristics: ■

The interface is not in any VLAN (not even VLAN 1).

■

The switch does not keep any Layer 2 switching table information for the interface.

■

Layer 3 settings, such as the IP address, are configured under the physical interface—just like a router.

■

The adjacency table lists the outgoing physical interface or PortChannel, which means that Layer 2 switching logic is not required in these cases.

Multilayer Switching

197

Ethernet PortChannels can be used as routed interfaces as well. To do so, as on physical routed interfaces, the no switchport command should be configured. (For PortChannels, the physical interfaces in the PortChannel must also be configured with the no switchport command.) Also, when using a PortChannel as a routed interface, PortChannel load balancing should be based on Layer 3 addresses because the Layer 2 addresses will mostly be the MAC addresses of the two MLS switches on either end of the PortChannel. PortChannels may also be used as Layer 2 interfaces when doing MLS. In that case, VLAN interfaces would be configured with IP address, and the PortChannel would simply act as any other Layer 2 interface. Table 6-5 lists some of the specifics about each type of Layer 3 interface. Table 6-5

MLS Layer 3 Interfaces Forwarding to Adjacent Device

Configuration Requirements

VLAN interface

Uses Layer 2 logic and L2 MAC address table

Create VLAN interface; VLAN must also exist

Physical (routed) interface

Forwards out physical interface

Use no switchport command to create a routed interface

PortChannel (switched) interface

Not applicable; just used as another Layer 2 forwarding path

No special configuration; useful in conjunction with VLAN interfaces

PortChannel (routed) interface

Balances across links in PortChannel

Needs no switchport command in order to be used as a routed interface; optionally change loadbalancing method

Interface

MLS Configuration The upcoming MLS configuration example is designed to show all of the configuration options. The network design is shown in Figures 6-4 and 6-5. In Figure 6-4, the physical topology is shown, with routed ports, VLAN trunks, a routed PortChannel, and access links. Figure 6-5 shows the same network, with a Layer 3 view of the subnets used in the network.

198

Chapter 6: IP Forwarding (Routing)

Figure 6-4

Physical Topology: Example Using MLS Server 1

FA0/1 FA0/0

Server 4

VLAN 11

MLS

R

T

SW1

R1

T

SW3

VLAN 12

SW4

VLAN 11

T R T FA0/0

R

SW2 MLS

R2

T

VLAN 12

Server 3

T — Trunk (also a switchport) R — Routed Port

Figure 6-5

Server 2

Layer 3 Topology View: Example Using MLS Server 1

Server 3 MLS

V11

172.31.21.0/24

R1

VLAN 11 172.31.11.0/24

SW1

V12

172.31.23.0/24 V11

VLAN 12

172.31.22.0/24

R2

SW2 MLS

V12

172.31.12.0/24

Server 4

Vx — VLAN Interface x

Server 2

A few design points bear discussion before jumping into the configuration. First, SW1 and SW2 need Layer 2 connectivity to support traffic in VLANs 11 and 12. In other words, you need a Layer 2 trunk between SW1 and SW2, and for several reasons. Focusing on the Layer 2 portions of the network on the right side of Figure 6-4, SW1 and SW2, both distribution MLS switches, connect to SW3 and SW4, which are access layer switches. SW1 and SW2 are responsible for providing

Multilayer Switching

199

full connectivity in VLANs 11 and 12. To fully take advantage of the redundant links, SW1 and SW2 need a Layer 2 path between each other. Additionally, this design uses SW1 and SW2 as Layer 3 switches, so the hosts in VLANs 11 and 12 will use SW1 or SW2 as their default gateway. For better availability, the two switches should use HSRP, VRRP, or GLBP. Regardless of which protocol is used, both SW1 and SW2 need to be in VLANs 11 and 12, with connectivity in those VLANs, to be effective as default gateways. In addition to a Layer 2 trunk between SW1 and SW2, to provide effective routing, it makes sense for SW1 and SW2 to have a routed path between each other as well. Certainly, SW1 needs to be able to route packets to router R1, and SW2 needs to be able to route packets to router R2. However, routing between SW1 and SW2 allows for easy convergence if R1 or R2 fails. Figure 6-4 shows two alternatives for routed connectivity between SW1 and SW2, and one option for Layer 2 connectivity. For Layer 2 connectivity, a VLAN trunk needs to be used between the two switches. Figure 6-4 shows a pair of trunks between SW1 and SW2 (labeled with a circled T) as a Layer 2 PortChannel. The PortChannel would support the VLAN 11 and 12 traffic. To support routed traffic, the figure shows two alternatives: simply route over the Layer 2 PortChannel using VLAN interfaces, or use a separate routed PortChannel. First, to use the Layer 2 PortChannel, SW1 and SW2 could simply configure VLAN interfaces in VLANs 11 and 12. The alternative configuration uses a second PortChannel that will be used as a routed PortChannel. However, the routed PortChannel does not function as a Layer 2 path between the switches, so the original Layer 2 PortChannel must still be used for Layer 2 connectivity. Upcoming Example 6-3 shows both configurations. Finally, a quick comment about PortChannels is needed. This design uses PortChannels between the switches, but they are not required. Most links between switches today use at least two links in a PortChannel, for the typical reasons—better availability, better convergence, and less STP overhead. This design includes the PortChannel to point out a small difference between the routed interface configuration and the routed PortChannel configuration. Example 6-3 shows the configuration for SW1, with some details on SW2. Example 6-3

MLS-Related Configuration on Switch1

! Below, note that the switch is in VTP transparent mode, and VLANs 11 and 12 are ! configured, as required. Also note the ip routing global command, without which ! the switch will not perform Layer 3 switching of IP packets. v lan 11 ! vlan 12 ! The ip routing global command is required before the MLS will perform ! Layer 3 forwarding. ip routing

continues

200

Chapter 6: IP Forwarding (Routing)

Example 6-3

MLS-Related Configuration on Switch1 (Continued)

! vtp doma in CCIE-domain vtp mod e transparent ! Next the no switchport command makes PortChannel a routed port. On a routed ! port, an IP address can be added to the interface. interfac e Port-channel1 no s w i t c h p o r t i p a d d r e s s 1 7 2 . 3 1 . 2 3 . 2 0 1 2 5 5 . 2 5 5 . 2 5 5 .0 ! Below, similar configuration on the interface connected to Router1. interface FastEt hernet0/1 no switchp ort i p a d d r e s s 1 7 2 . 3 1 . 2 1 . 2 0 1 2 5 5 . 2 5 5 . 2 55 . 0 ! Next, the configuration shows basic PortChannel commands, with the ! no switchport command being required due to the same command on PortChannel. interface GigabitEthernet0/1 no switchport no ip address c h a n n e l - g r o u p 1 m od e d e s i r a b l e ! i n t e r f a c e G i ga b i t E t h e r n e t 0 / 2 n o s w i t c hp o r t no ip address channel-group 1 mode desirable ! Next, interface VLAN 11 gives Switch1 an IP presence in VLAN11. Devices in VLAN ! 11 can use 172.31.11.201 as their default gateway. However, using HSRP is ! better, so Switch1 has been configured to be HSRP primary in VLAN11, and Switch2 ! to be primary in VLAN12, with tracking so that if Switch1 loses its connection ! to Router1, HSRP will fail over to Switch2. interface Vlan 11 ip address 172.31.11.201 255.255.255.0 no ip red irects s t an d b y 1 1 i p 1 7 2 . 3 1 . 1 1 . 2 5 4 st a n d b y 1 1 p r i o r i t y 9 0 standby 11 track FastEthernet0/1 ! Below, VLAN12 has similar configuration settings, but with a higher (better) ! HSRP priority than Switch2’s VLAN 12 interface. interface Vlan12 ip address 172.31.12.201 255.255.255.0 no ip redire cts s t an d b y 1 2 i p 1 7 2 . 3 1 . 1 2 . 2 5 4 st a n d b y 1 2 p r i o r i t y 1 1 0 standby 12 track FastEthernet0/1

Policy Routing

201

NOTE For MLS switches to route using VLAN interfaces, two other actions are required: The corresponding VLANs must be created, and the ip routing global command must have been configured. (MLS switches will not perform Layer 3 routing without the ip routing command, which is not enabled by default.) If the VLAN interface is created before either of those actions, the VLAN interface sits in a “down and down” state. If the VLAN is created next, the VLAN interface is in an “up and down” state. Finally, after adding the ip routing command, the interface is in an “up and up” state. As stated earlier, the routed PortChannel is not required in this topology. It was included to show an example of the configuration, and to provide a backdrop from which to discuss the differences. However, as configured, SW1 and SW2 are Layer 3 adjacent over the routed PortChannel as well as via their VLAN 11 and 12 interfaces. So, they could exchange IGP routing updates over three separate subnets. In such a design, the routed PortChannel was probably added so that it would be the normal Layer 3 path between SW1 and SW2; care should be taken to tune the IGP implementation so that this route is chosen instead of the routes over the VLAN interfaces.

Policy Routing All the options for IP forwarding (routing) in this chapter had one thing in common: The destination IP address in the packet header was the only thing in the packet that was used to determine how the packet was forwarded. Policy routing allows a router to make routing decisions based on information besides the destination IP address. Policy routing’s logic begins with the ip policy command on an interface. This command tells IOS to process incoming packets with different logic before the normal forwarding logic takes place. (To be specific, policy routing intercepts the packet after Step 2, but before Step 3, in the routing process shown in Figure 6-1.) IOS compares the received packets using the route map referenced in the ip policy command. Figure 6-6 shows the basic logic. Specifying the matching criteria for policy routing is relatively simple compared to defining the routing instructions using the set command. The route maps used by policy routing must match either based on referring to an ACL (numbered or named IP ACL, using the match ip address command) or based on packet length (using the match length command). To specify the routing instructions—in other words, where to forward the packet next—the set command is used. Table 6-6 lists the set commands, and provides some insight into their differences.

202

Chapter 6: IP Forwarding (Routing)

Figure 6-6

Basic Policy Routing Logic Policy Routing Enabled?

Yes

Match 1st Clause?

Yes

No

No

Match 2nd Clause?

Yes

No

Match Last Clause?

Permit or Deny?

Permit

Route Based on set Command

Deny

Yes

Destination Based Routing (Normal)

No

Table 6-6

Policy Routing Instructions (set Commands) Command

Comments

set ip next-hop ip-address [. . . ip-address]

Next-hop addresses must be in a connected subnet; forwards to the first address in the list for which the associated interface is up.

set ip default next-hop ip-address [. . . ip-address]

Same logic as previous command, except policy routing first attempts to route based on the routing table.

set interface interface-type interface-number [. . . interfacetype interface-number]

Forwards packets using the first interface in the list that is up.

set default interface interface-type interface-number [. . . interfacetype interface-number]

Same logic as previous command, except policy routing first attempts to route based on the routing table.

set ip precedence number | name

Sets IP precedence bits; can be decimal value or ASCII name.

set ip tos [number]

Sets entire ToS byte; numeric value is in decimal.

Policy Routing

203

The first four set commands in Table 6-6 are the most important ones to consider. Essentially, you set either the next-hop IP address or the outgoing interface. Use the outgoing interface option only when it is unambiguous—for instance, do not refer to a LAN interface or multipoint Frame Relay subinterface. Most importantly, note the behavior of the default keyword in the set commands. Use of the default keyword essentially means that policy routing tries the default (destination based) routing first, and resorts to using the set command details only when the router finds no matching route in the routing table. The remaining set commands set the bits inside the ToS byte of the packet; refer to Chapter 12, “Classification and Marking,” for more information about the ToS byte and QoS settings. Note that you can have multiple set commands in the same route-map clause. For instance, you may want to define the next-hop IP address and mark the packet’s ToS at the same time. Figure 6-7 shows a variation on the same network used earlier in this chapter. Router3 and Router4 are now at the same site, connected to the same LAN, and each has PVCs connecting to Router1 and Router2. Figure 6-7

Policy Routing Example on Router3 172.31.0.0, /24 Prefixes

Client 3 103.3 E0/0

S0/0.134

R3

R1 Telnet Server–

GW is 103.3

172.31.11.201

S0/0.32

172.31.123.2

103.4

R4 DLCI 400

R2

Example 6-4 shows three separate policy routing configurations on Router3. The first configuration forwards Telnet traffic over the PVC to Router2 (next hop 172.31.123.2). The next configuration does the same thing, but this time using the set interface command. The final option shows a nonworking case with Router3 specifying its LAN interface as an outgoing interface. Example 6-4

Policy Routing Example on Router3

! Below, Router3 is configured with three route maps, one of which is enabled on ! interface e0/0 with the ip policy route-map to-R2-nexthop command. The two ! route maps that are not referenced in the ip policy command are used ! later in the configuration. Router3# sh run ! Lines omitted for brevity interface Ethernet0/0 mac-address 0200.3333.3333 ip address 172.31.103.3 255.255.255.0

continues

204

Chapter 6: IP Forwarding (Routing)

Example 6-4

Policy Routing Example on Router3 (Continued)

ip policy route-map to-R2-nexthop ! interface Serial0/0.32 point-to-point ip address 172.31.123.3 255.255.255.0 frame-relay interface-dlci 200 ! interface Serial0/0.134 point-to-point ip address 172.31.134.3 255.255.255.0 frame-relay interface-dlci 100 ! access-list 111 permit tcp any any eq telnet ! This route-map matches all telnet, and picks a route through R2. route-map to-R2-nexthop permit 10 match ip address 111 set ip next-hop 172.31.123.2 ! This route-map matches all telnet, and picks a route out E0/0. route-map to-R4-outgoing permit 10 match ip address 111 set interface Ethernet0/0 ! This route-map matches all telnet, and picks a route out S0/0.32. route-map to-R2-outgoing permit 10 match ip address 111 set interface Serial0/0.32 ! debugging is enabled to prove policy routing is working on Router3. Router3# debug ip policy Policy routing debugging is on ! Not shown, a Client3 tries to telnet to 172.31.11.201 ! Below, a sample of the debug messages created for a single policy-routed packet. 06:21:57: IP: route map to-R2-nexthop, item 10, permit 06:21:57: IP: Ethernet0/0 to Serial0/0.32 172.31.123.2 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! ! Next, Router3 uses a different route-map. This one sets the outgoing interface to ! S0/0.32.

The Outgoing interface option works, because it is a point-to-point

! subinterface Router3# conf t Enter configuration commands, one per line.

End with CNTL/Z.

Router3(config)# int e 0/0 Router3(config-if)# ip policy route-map to-R2-outgoing Router3(config-if)# ^Z ! Not shown, the same user with default gateway of Router3 tries to telnet again. ! Below, the sample debug messages are identical as the previous set of messages. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 06:40:51: IP: route map to-R2-outgoing, item 10, permit 06:40:51: IP: Ethernet0/0 to Serial0/0.32 172.31.123.2 ! Next, switching to a third route-map that sets the outgoing interface to E0/0. Router3# conf t Enter configuration commands, one per line.

End with CNTL/Z.

Policy Routing

Example 6-4

205

Policy Routing Example on Router3 (Continued)

Router3(config)# int e 0/0 Router3(config-if)# ip policy route-map to-R4-outgoing Router3(config-if)# ^Z ! Not shown, the same user with default gateway of Router3 tries to telnet again. ! Router3 actually sends an ARP request out e0/0, looking for ! the IP address in the destination of the packet – 172.31.11.201, the address ! to which the user is telnetting.

Also below, Router3 shows that the ARP table

! entry for 172.31.11.201 is incomplete. Router3# sh ip arp Protocol

Address

Internet

172.31.11.201

Age (min)

Internet

172.31.104.3

−

0200.3333.3333

ARPA

Ethernet0/0

Internet

172.31.104.4

0

0200.4444.4444

ARPA

Ethernet0/0

0

Hardware Addr

Type

Incomplete

ARPA

Interface

The first two route maps in the example were relatively simple, with the last route map showing why specifying a multi-access outgoing interface is problematic. In the first two cases, the telnet works fine; to verify that it was working, the debug ip policy command was required. The third route map (to-R4-outgoing) sets the output interface to Router3’s E0/0 interface. Because Router3 does not have an associated next-hop IP address, Router3 sends an ARP request asking for 172.31.11.201’s MAC address. As shown in the show ip arp command output, Router3 never completes its ARP entry. To work around the problem, assuming that the goal is to forward the packets to Router4 next, the configuration in Router3 should refer to the next-hop IP address instead of the outgoing interface E0/0. NOTE Policy Routing for this particular topology fails due to a couple of tricky side effects of ARP. At first glance, you might think that the only thing required to make the to-R4-outgoing policy work is for R4 to enable proxy ARP. In fact, if R4 is then configured with an ip proxyarp interface subcommand, R4 does indeed reply to R3’s ARP for 172.31.11.201. R4 lists its own MAC address in the ARP reply. However, R3 rejects the ARP reply, because of a basic check performed on ARPs. R3’s only IP route matching address 172.31.11.201 points over the WAN interface, and routers check ARP replies to make sure they list a sensible interface. From R3’s perspective, the only sensible interface is one through which the destination might possibly be reached. So, R3’s logic dictates that it should never hear an ARP reply regarding 172.31.11.201 coming in its fa0/0 interface, so R3 rejects the (proxy) ARP reply from R4. To see all of this working in a lab, re-create the topology, and use the debug ip arp and debug policy commands.

206

Chapter 6: IP Forwarding (Routing)

Optimized Edge Routing and Performance Routing Performance routing (PfR) and optimized edge routing (OER) were added to the CCIE R&S qualification exam blueprint in 2009. OER came first, and Cisco has extended its functionality and renamed it PfR. This approach is similar to Cisco’s change of nomenclature from Service Assurance Agent (SAA) to IP service level agreement (IP SLA). (Incidentally, PfR uses the IP SLA feature internally.) As you’ll see in the example later in this section, PfR uses commands that begin with oer for configuration. According to the Cisco documentation, the OER configuration commands may eventually be replaced and deprecated, but you should concentrate on the concepts of PfR and the oer commands as you learn this material. Throughout this section, we refer to this feature as PfR for clarity; but first, here’s a brief look at how OER came into existence. OER was created to extend the capability of routers to more optimally route traffic than routing protocols can provide on their own. To do this, OER takes into account the following information: ■

Packet loss

■

Response time

■

Path availability

■

Traffic load distribution

By adding this information into the routing decision process, OER can influence routing to avoid links with unacceptable latency, packet loss, or network problems severe enough to cause serious application performance problems, but not severe enough to trigger routing changes by the routing protocols in use. Furthermore, taking into account that many modern networks use multiple service provider circuits and typically do little or no load balancing between them, OER provides for this functionality. OER performs these functions automatically, but also allows network administrators to manually configure them in a highly granular way if desired. OER uses a five-phase operational model: ■

Profile—Learn the flows of traffic that have high latency or high throughput.

■

Measure—Passively/actively collect traffic performance metrics.

■

Apply policy—Create low and high thresholds to define in-policy and out-of-policy (OOP) performance categories.

■

Control—Influence traffic by manipulating routing or by routing in conjunction with PBR.

■

Verify—Measure OOP event performance and adjust policy to bring performance in-policy.

OER and PfR influence traffic by collecting information (more on that later) and then injecting new routes into the routing table with the appropriate routing information, tags, and other

Optimized Edge Routing and Performance Routing

207

attributes (for BGP routes) to steer traffic in a desired direction. The new routes are redistributed into the IGP. As conditions change, these new routes may be removed, or more may be added. To provide for the required level of granularity, OER and PfR can split up subnets or extract part of a subnet or prefix from the remainder of that prefix by injecting a longer match into the routing table. Because the longest match is the first criteria in a Cisco router’s decision-making process about where to send traffic, OER and PfR don’t require any deep changes in how routers make decisions. You can think of this feature set as providing more information to the router to help it make better routing decisions, on a flow-by-flow basis. PfR officially stands for performance routing, but it’s also known as protocol-independent routing optimization, or PIRO. From this point on, this section refers to this feature solely as PfR. PfR learns about network performance using the IP SLA and NetFlow features (one or both) in Cisco IOS. PfR has the following requirements and conditions: ■

CEF must be enabled.

■

IGP/BGP routing must be configured and working.

■

PfR does not support Multiprotocol Label Switching Provider Edge - Customer (MPLS PECE) or any traffic within the MPLS network, because PfR does not recognize MPLS headers.

■

PfR uses redistribution of static routes into the routing table, with a tag to facilitate control.

PfR extends beyond OER’s original capability by providing routing optimization based on traffic type, through application awareness. PfR lets a router select the best path across a network based on the application traffic requirements. For example, voice traffic requires low latency, low jitter, and low error rates. These are the fundamental attributes of PfR: ■

Optimizes traffic path based on application type, performance requirements, and network performance

■

Controls outbound traffic using redistributed IGP routes

■

Controls inbound traffic (as of Cisco IOS 12.4(9)T) in BGP networks by prepending autonomous systems or through BGP communities on selected BGP Network Layer Reachability Information (NLRI), causing the neighboring BGP routers to take this information into account

■

Provides logical link bundling

■

Can performs passive monitoring using the Cisco IOS NetFlow feature

■

Can perform active monitoring using the Cisco IOS IP SLA feature

208

Chapter 6: IP Forwarding (Routing)

■

Can perform active and passive monitoring simultaneously

■

Can operate in monitor-only mode to collect information that helps network administrators determine the benefit of implementing PfR

■

Performs dynamic load balancing

■

Can reroute traffic in as little as 3 seconds

■

Performs automatic path optimization

■

Offers “good” mode: finds an alternative route when a defined threshold is exceeded

■

Offers “best” mode: always switches traffic to the route with the best performance

■

Supports robust reporting for traffic analysis and path assessment and troubleshooting purposes

■

Can split prefixes in the routing table to provide differentiated routing for a single host or a subset of hosts compared to the prefixes in the original routing table

■

Allows prioritization of its decision criteria to arbitrate conflicts in applying policy.

PfR is supported across the Cisco 1800–3800 series ISR platforms and the ASR, 7200, and 7600 routers, and the Catalyst 6500 series. It requires the SP Services, Advanced IP Services, Enterprise Services, or Advanced Enterprise Services feature set in Cisco IOS.

Device Roles in PfR To deploy PfR, you must designate the router’s role in the PfR environment. The two roles are ■

Master Controller (MC)—Configured using the oer master command, enables PfR and configures this router to be the decision maker in a cluster of PfR routers (typically 10 or fewer routers). The MC learns specified information from the BRs and makes configuration decisions for the network based on this information.

■

Border Router (BR)—Configured using the oer border command, this mode makes a router subordinate to the MC. The BR accepts commands from the MC and provides information to the MC in its role in the PfR environment.

A single router can act as both the MC and a BR. Typically, you’d configure a BR to be the MC if there’s one edge router in the environment and it has two or more external WAN or Internet connections. This is common in small network environments where there’s a primary Internet connection and a backup. In larger deployments where multiple edge routers are present and each one connects to one or more external links, one of the BR routers could also act as the MC, or the MC could be a different, internal router. In a case where two edge routers, each with a WAN or

Optimized Edge Routing and Performance Routing

209

Internet link, are present in the environment, choose one of them to be the MC. Both will also be configured as BRs.

MC High Availability and Failure Considerations BR and MC routers maintain communication with each other using keepalives. If keepalives from the MC stop, the BRs remove any PfR-added routing information and the network returns to its pre-PfR configuration. For PfR high availability, you can configure more than one MC in a PfR deployment. PfR traffic classes can be defined by IP address, protocol, port number, or differentiated services code point (DSCP) setting. This is useful if you have specific traffic flows for which you want to manually configure PfR optimization. In active mode, PfR uses the IP SLA feature. BRs source probes to the MC for delay, jitter, reachability, or mean opinion score (MOS). A MOS is calculated using voice-like packets generated using the IP SLA feature to measure jitter, latency, and packet loss. In passive mode, PfR uses NetFlow information based on traffic classes to make decisions. To take advantage of PfR for iBGP routes, you must redistribute them into the IGP. PfR doesn’t directly support iBGP.

PfR Configuration Example 6-5 shows the configuration of a PfR MC router (PfR-MC) and two PfR BRs (PfR-BREast and PfR-BR-West). The goals for PfR in this configuration are as follows: ■

The MC (172.17.10.1) will learn prefixes with the longest delay.

■

The monitoring period is 5 minutes.

■

During each monitoring period, the MC will track up to 200 prefixes.

■

The time between monitoring periods is 15 minutes.

■

The two BRs (172.17.100.1 and 172.17.104.1) each have an inside Ethernet interface and an outside serial interface. In the context of PfR, each router needs to understand which interfaces are on the inside and which are on the outside, much like NAT.

■

Each BR has a serial interface with different bandwidth and operating cost. The MC will take these into account in its activities.

■

The BRs will source Active probes from their Fa0/0 interfaces.

210

Chapter 6: IP Forwarding (Routing)

Example 6-5

Configuring PfR

! On the Master Controller: PfR-MC# config term PfR-MC(config)# key-chain key1 PfR-MC(config-keychain)# key 1 PfR-MC(config-keychain-key)# key-string pfr PfR-MC(config-keychain-key)# exit PfR-MC(config-keychain)# exit PfR-MC(config)# oer master PfR-MC(config-oer-mc)# logging PfR-MC(config-oer-mc)# mode route control PfR-MC(config-oer-mc)# max prefix total 1000 PfR-MC(config-oer-mc)# backoff 90 3000 300 PfR-MC(config-oer-mc)# learn PfR-MC(config-oer-mc-learn)# delay PfR-MC(config-oer-mc-learn)# monitor period 5 PfR-MC(config-oer-mc-learn)# periodic interval 15 PfR-MC(config-oer-mc-learn)# exit PfR-MC(config-oer-mc)# border 172.17.100.1 key-chain key1 PfR-MC(config-oer-mc-br)# interface fa0/0 internal PfR-MC(config-oer-mc-br-if)# exit PfR-MC(config-oer-mc-br)# interface serial0/0 external PfR-MC(config-oer-mc-br-if)# max-xmit-utilization absolute 1500 PfR-MC(config-oer-mc-br-if)# cost-minimization fixed fee 1000 PfR-MC(config-oer-mc-br-if)# border 172.17.104.1 key-chain key1 PfR-MC(config-oer-mc-br)# interface fa0/0 internal PfR-MC(config-oer-mc-br)# interface serial0/0 external PfR-MC(config-oer-mc-br-if)# max-xmit-utilization absolute 1000 PfR-MC(config-oer-mc-br-if)# cost-minimization fixed fee 800 PfR-MC(config-oer-learn)# end PfR-MC# ! Now to Border Router 1: PfR-BR-East# config term PfR-BR-East(config)# key-chain key1 PfR-BR-East(config-keychain)# key 1 PfR-BR-East(config-keychain-key)# key-string pfr PfR-BR-East(config-keychain-key)# exit PfR-BR-East(config-keychain)# exit PfR-BR-East(config)# oer border PfR-BR-East(config-oer-br)# master 172.17.10.1 key-chain key1 PfR-BR-East(config-oer-br)# local fa0/0 PfR-BR-East(config-oer-br)# active-probe address source interface FastEthernet0/0 PfR-BR-East(config-oer-br)# end PfR-BR-East# ! Now to Border Router 2: PfR-BR-West# config term PfR-BR-West(config)# key-chain key1 PfR-BR-West(config-keychain)# key 1

Optimized Edge Routing and Performance Routing

Example 6-5

211

Configuring PfR (Continued)

PfR-BR-West(config-keychain-key)# key-string pfr PfR-BR-West(config-keychain-key)# exit exit PfR-BR-West(config-keychain)#e PfR-BR-West(config)# oer border PfR-BR-West(config-oer-br)# master 172.17.10.1 key-chain key1 PfR-BR-West(config-oer-br)# local fa0/0 PfR-BR-West(config-oer-br)# active-probe address source interface FastEthernet0/0 PfR-BR-West(config-oer-br)# end

Some of the key show commands for PfR include show oer border routes, show oer master border, show oer master learn list, show oer master prefix, show oer master traffic-list, and show oer master policy. As you will see as you explore PfR more deeply, it is remarkably powerful. As you would expect, it also has many configuration options. For the CCIE R&S qualifying exam, you should understand the concepts of PfR and how it operates and its core functionality.

GRE Tunnels Generic routing encapsulation (GRE) defines a method of tunneling data from one router to another. To tunnel the traffic, the sending router encapsulates packets of one networking protocol, called the passenger protocol, inside packets of another protocol, called the transport protocol, transporting these packets to another router. The receiving router de-encapsulates and forwards the original passenger protocol packets. This process allows the routers in the network to forward traffic that might not be natively supported by the intervening routers. For instance, if some routers did not support IP multicast, the IP multicast traffic could be tunneled from one router to another using IP unicast packets. Generic routing encapsulation, or GRE, tunnels are useful for a variety of tasks. From the network standpoint, GRE tunnel traffic is considered GRE; that is, it’s not IP unicast or multicast or IPsec or whatever else is being encapsulated. Therefore, you can use GRE to tunnel traffic that might not otherwise be able to traverse the network. An example of this is encapsulating multicast IP traffic within a GRE tunnel to allow it to pass across a network that does not support multicast. GRE tunnels can also be used to encapsulate traffic so that the traffic inside the tunnel is unaware of the network topology. Regardless of the number of network hops between the source and destination, the traffic passing over the tunnel sees it as a single hop. Because tunnels hide the network topology, the actual traffic path across a network is also unimportant to the traffic inside the tunnel. If loopback addresses are used for source and destination addresses, the tunnel provides connectivity between the source and destination as long as there is any available route between the

212

Chapter 6: IP Forwarding (Routing)

loopbacks. Even if the normal egress interface were to go down, traffic across the tunnel could continue to flow. The router must be configured to pass the desired traffic across the tunnel. This is often accomplished using a static route to point traffic to the tunnel interface. Example 6-6 shows a typical tunnel configuration on two routers. Both routers source traffic from a loopback interface and point to the corresponding loopback interface. IP addresses in a new subnet on these routers are assigned to the loopback interfaces. Example 6-6

GRE Tunnel Configuration

R2# show run int lo0 interface Loopback0 ip address 150.1.2.2 255.255.255.0 R2# show run int tun0 interface Tunnel0 ip address 192.168.201.2 255.255.255.0 tunnel source Loopback0 tunnel destination 150.1.3.3 ! Now on to R3: R3# show run int lo0 interface Loopback0 ip address 150.1.3.3 255.255.255.128 R3# show run int tun0 interface Tunnel0 ip address 192.168.201.3 255.255.255.0 tunnel source Loopback0 tunnel destination 150.1.2.2 R3# show ip interface brief Interface

IP-Address

OK? Method Status

Protocol

Serial0/2

144.254.254.3

YES TFTP

up

up

Serial0/3

unassigned

YES NVRAM

up

down

Virtual-Access1

unassigned

YES unset

up

up

Loopback0

150.1.3.3

YES NVRAM

up

up

Tunnel0

192.168.201.3

YES manual up

up

Foundation Summary

213

Foundation Summary This section lists additional details and facts to round out the coverage of the topics in this chapter. Unlike most of the Cisco Press Exam Certification Guides, this “Foundation Summary” does not repeat information presented in the “Foundation Topics” section of the chapter. Please take the time to read and study the details in the “Foundation Topics” section of the chapter, as well as review items noted with a Key Topic icon. Table 6-7 lists the protocols mentioned in or pertinent to this chapter and their respective standards documents. Table 6-7

Protocols and Standards for Chapter 6 Name

Standardized In

Address Resolution Protocol (ARP)

RFC 826

Reverse Address Resolution Protocol (RARP)

RFC 903

Frame Relay Inverse ARP (InARP)

RFC 2390

Frame Relay Multiprotocol Encapsulation

RFC 2427

Differentiated Services Code Point (DSCP)

RFC 2474

Table 6-8 lists some of the key IOS commands related to the topics in this chapter. (The command syntax for switch commands was taken from the Catalyst 3560 Multilayer Switch Command Reference, 12.2(25)SEE.) Router-specific commands were taken from the IOS 12.3 mainline command reference.) Table 6-8

Command Reference for Chapter 6 Command

Description

[no] ip classless

Enables classless (ip classless) or classful (no ip classless) forwarding.

show ip arp

EXEC command that displays the contents of the IP ARP cache.

show frame-relay map

Router exec command that lists the mapping information between Frame Relay DLCIs and Layer 3 addresses.

frame-relay interface-dlci

Configuration command that associates a particular DLCI with a subinterface. continues

214

Chapter 6: IP Forwarding (Routing)

Table 6-8

Command Reference for Chapter 6 (Continued) Command

Description

[no] switchport

Switch interface subcommand that toggles an interface between a Layer 2 switched function (switchport) and a routed port (no switchport).

clear frame-relay inarp

Router exec command that clears all InARP-learned entries from the Frame Relay mapping table.

[no] ip route-cache cef

Interface subcommand that enables or disables CEF switching on an interface.

[no] ip cef

Global configuration command to enable (or disable) CEF on all interfaces.

debug frame-relay events

Displays messages about various events, including InARP messages.

show frame-relay map

Displays information about Layer 3 to Layer 2 mapping with Frame Relay.

frame-relay map protocol protocol-address {dlci} [broadcast] [ietf | cisco]

Interface subcommand that maps a Layer 3 address to a DLCI.

[no] frame-relay inversearp [protocol] [dlci]

Interface subcommand that enables or disables InARP.

[no] ip routing

Enables IP routing; defaults to no ip routing on a multilayer switch.

ip policy route-map map-tag

Router interface subcommand that enables policy routing for the packets entering the interface.

oer master

Configures a router for PfR in Master Controller mode.

oer border

Configures a router for PfR in Border Controller mode.

learn

Configures a PfR MC to learn information based on specified criteria, including prefix length, delay, protocol, throughput, and BGP and nonBGP prefix type.

show oer border routes [static | bgp]

Displays information on routes controlled by OER on an OER Border Router.

show oer master prefix [detail | inside [detail] | learned [delay | inside | throughput] | prefix [detail | policy | report | traceroute [exit-id | border-address | current] [now]]]

Displays the status of prefixes monitored in OER or PfR operation.

show oer master policy [sequence-number | policyname | default | dynamic]

Entered on an OER master router, this command displays the default policy and the policies applied to an OER map.

Memory Builders

Table 6-8

215

Command Reference for Chapter 6 (Continued) Command

Description

interface tunnel number

Enters tunnel configuration mode, specifies a tunnel interface number.

tunnel source source-interface

Specifies a source interface for the tunnel.

tunnel destination ip-address

Specifies the destination IP address for the tunnel.

tunnel mode [gre {ip | multipoint} | dvmrp | ipip | mpls | nos]

The default tunnel mode is GRE IP.

Refer to Table 6-6 for the list of set commands related to policy routing.

Memory Builders The CCIE Routing and Switching written exam, like all Cisco CCIE written exams, covers a fairly broad set of topics. This section provides some basic tools to help you exercise your memory about some of the broader topics covered in this chapter.

Fill In Key Tables from Memory Appendix G, “Key Tables for CCIE Study,” on the CD in the back of this book contains empty sets of some of the key summary tables in each chapter. Print Appendix G, refer to this chapter’s tables in it, and fill in the tables from memory. Refer to Appendix H, “Solutions for Key Tables for CCIE Study,” on the CD to check your answers.

Definitions Next, take a few moments to write down the definitions for the following terms: policy routing, process switching, CEF, MLS, ARP, proxy ARP, routed interface, InARP, fast switching, TTL, classless routing, classful routing, FIB, adjacency table, control plane, switched interface, data plane, IP routing, IP forwarding Refer to the glossary to check your answers.

Further Reading For a good reference on load balancing with CEF, refer to http://cisco.com/en/US/partner/tech/ tk827/tk831/technologies_tech_note09186a0080094806.shtml. This website requires a cisco.com account. For details on OER and PfR, visit http://www.cisco.com/go/oer and http://www.cisco.com/go/pfr.

Blueprint topics covered in this chapter: This chapter covers the following subtopics from the Cisco CCIE Routing and Switching written exam blueprint. Refer to the full blueprint in Table I-1 in the Introduction for more details on the topics covered in each chapter and their context within the blueprint. ■

Best Path

■

Loop Free Paths

■

EIGRP Operations when Alternate Loop-free Paths Are Available and when No Alternate Loop-free Paths Are Available

■

EIGRP Queries

■

Manual Summarization

■

Autosummarization

■

EIGRP Stubs

CHAPTER

7

EIGRP This chapter covers most of the features, concepts, and commands related to EIGRP. Chapter 9, “IGP Route Redistribution, Route Summarization, Default Routing, and Troubleshooting,” covers a few other details of EIGRP—in particular, route redistribution, route filtering when redistributing, and route summarization.

“Do I Know This Already?” Quiz Table 7-1 outlines the major headings in this chapter and the corresponding “Do I Know This Already?” quiz questions. Table 7-1

“Do I Know This Already?” Foundation Topics Section-to-Question Mapping Foundation Topics Section

Questions Covered in This Section

EIGRP Basics and Steady-State Operation

1–4

EIGRP Convergence

5–7

EIGRP Configuration

8–9

Score

Total Score

In order to best use this pre-chapter assessment, remember to score yourself strictly. You can find the answers in Appendix A, “Answers to the ‘Do I Know This Already?’ Quizzes.” 1.

2.

Which of the following items are true of EIGRP? a.

Authentication can be done using MD5 or clear text.

b.

Uses UDP port 88.

c.

Sends full or partial updates as needed.

d.

Multicasts updates to 224.0.0.10.

Four routers (R1, R2, R3, and R4) are attached to the same VLAN. R1 has been configured for an EIGRP Hello timer of 3. R2 has been configured with a metric weights 0 0 0 1 0 0 command. R3 has been configured with a hold time of 11 seconds. Their IP addresses are

218

Chapter 7: EIGRP

10.1.1.1, 10.1.1.2, 10.1.1.3, and 10.1.1.4, with /24 prefixes, except R4, which has a /23 prefix configured. All other related parameters are set to their default. Select the routers that are able to collectively form neighbor relationships.

3.

a.

R1

b.

R2

c.

R3

d.

R4

e.

None of them can form a neighbor relationship.

In the following command output, what do the numbers in the column labeled “H” represent? R1# show ip eigrp neighbors IP-EIGRP neighbors for process 1 H Address Interface 2 1 0

4.

5.

172.31.11.2 172.31.11.202 172.31.11.201

Fa0/0 Fa0/0 Fa0/0

Hold Uptime (sec) 4 00:03:10 11 00:04:43 14 00:05:11

a.

The current Hold Time countdown

b.

The number of seconds before a Hello is expected

c.

The order in which the neighbors came up

d.

None of the other answers is correct

SRTT (ms) 1 1 1927

RTO 4500 4500 5000

Q Cnt 0 0 0

Seq Num 233 81 84

Which of the following is not true regarding the EIGRP Update message? a.

Updates require an acknowledgement with an Ack message.

b.

Updates can be sent to multicast address 224.0.0.10.

c.

Updates are sent as unicasts when they are retransmitted.

d.

Updates always include all routes known by a router, with partial routing information distributed as part of the EIGRP Reply message.

The output of a show ip eigrp topology command lists information about subnet 10.1.1.0/24, with two successors, and three routes listed on lines beginning with “via.” How many feasible successor routes exist for 10.1.1.0/24? a.

0

b.

1

c.

2

d.

3

e.

Cannot determine from the information given

“Do I Know This Already?” Quiz

6.

219

The following command output shows R11’s topology information for subnet 10.1.1.0/24. Then R11 and R12 (IP address 10.1.11.2) are connected to the same LAN segment. Then R11’s EIGRP Hold Time expires for neighbor R12. Which of the following is true about R11’s first reaction to the loss of its neighbor R12? R11# show ip eigrp topology ! lines omitted for brevity P 10.1.1.0/24, 1 successors, FD is 1456 via 10.1.11.2 (1456/1024), FastEthernet0/0 via 10.1.14.2 (1756/1424), Serial0/0.4

7.

8.

a.

R11 sends Updates to all neighbors poisoning its route to 10.1.1.0/24.

b.

R11 replaces the old route through 10.1.11.2 with the feasible successor route through 10.1.14.2.

c.

R11 sends Query messages to all other neighbors to ensure that the alternate route through 10.1.14.2 is loop free, before using the route.

d.

R11 first Queries only neighbors on interface fa0/0 for alternative routes before Querying the rest of its neighbors.

EIGRP router R11 has just changed its route to subnet 10.1.2.0/24 to the active state, and has sent a Query to five neighbors. Which of the following is true about the next step taken by R11? a.

R11 adds a new route to 10.1.2.0/24 to the routing table as soon as it receives an EIGRP Reply that describes a new route to 10.1.2.0/24.

b.

R11 can add a new route to 10.1.2.0/24 after receiving Reply messages from all 5 neighbors.

c.

R11 can add a new route for 10.1.2.0/24 to the routing table, even without five Reply messages, once the Hold timer expires.

d.

R11 can add a new route for 10.1.2.0/24 to the routing table, even without five Reply messages, once the Dead timer expires.

EIGRP router R11 has five interfaces, with IP address 10.1.1.11/24 (interface fa0/0), 10.1.2.11/24, 10.1.3.11/24, 10.1.4.11/24, and 10.1.5.11/24. Its EIGRP configuration is shown below. Which of the following answers is true regarding this router? router eigrp 1 network 10.1.0.0 0.0.3.255 passive-interface fa0/0

a.

R11 will send EIGRP Updates out fa0/0, but not process received EIGRP Updates.

b.

R11 will advertise connected subnets 10.1.3.0/24 and 10.1.4.0/24.

c.

R11 will advertise subnets 10.1.1.0/24 and 10.1.2.0/24.

d.

The network command does not match any interfaces, so EIGRP will essentially do nothing.

220

Chapter 7: EIGRP

9.

EIGRP router Br1 is a branch router with two Frame Relay subinterfaces (s0/0.1 and s0/0.2) connecting it to distribution routers. It also has one LAN interface, fa0/0. No other routers connect to the Br1 LAN. Which of the following scenarios prevent router Br1 from sending EIGRP Hellos out its fa0/0 interface? a.

The inclusion of the passive-interface fa0/0 command on Br1

b.

The inclusion of the eigrp stub command on Br1

c.

The inclusion of the eigrp stub receive-only command on Br1

d.

The lack of a network command that matches the IP address of Br1’s fa0/0 interface

EIGRP Basics and Steady-State Operation

221

Foundation Topics

EIGRP Basics and Steady-State Operation Many CCIE candidates have learned many of the details of EIGRP operation and configuration. EIGRP is widely deployed and is thoroughly covered on the CCNP BSCI exam. With that in mind, this chapter strives to review the key terms and concepts briefly, and get right to specific examples that detail EIGRP operation on a Cisco router. To that end, the chapter begins with Table 7-2, which lists some of the key features related to EIGRP. Table 7-2

EIGRP Feature Summary Feature

Description

Transport

IP, protocol type 88 (does not use UDP or TCP).

Metric

Based on constrained bandwidth and cumulative delay by default, and optionally load, reliability, and MTU.

Hello interval

Interval at which a router sends EIGRP Hello messages on an interface.

Hold timer

Timer used to determine when a neighboring router has failed, based on a router not receiving any EIGRP messages, including Hellos, in this timer period.

Update destination address

Normally sent to 224.0.0.10, with retransmissions being sent to each neighbor’s unicast IP address.

Full or partial updates

Full updates are used when new neighbors are discovered; otherwise, partial updates are used.

Authentication

Supports MD5 authentication only.

VLSM/classless

EIGRP includes the mask with each route, also allowing it to support discontiguous networks and VLSM.

Route Tags

Allows EIGRP to tag routes as they are redistributed into EIGRP.

Next-hop field

Supports the advertisement of routes with a different next-hop router than the advertising router.

Manual route summarization

Allows route summarization at any point in the EIGRP network.

Multiprotocol

Supports the advertisement of IPX and AppleTalk routes.

Hellos, Neighbors, and Adjacencies After a router has been configured for EIGRP, and its interfaces come up, it attempts to find neighbors by sending EIGRP Hellos (destination 224.0.0.10). Once a pair of routers have heard each other say Hello, they become adjacent—assuming several key conditions are met. Once neighbors pass the checks in the following list, they are considered to be adjacent. At that point,

222

Chapter 7: EIGRP

they can exchange routes and are listed in the output of the show ip eigrp neighbor command. Neighbors should always form when these conditions are met, regardless of link type. ■

Must pass the authentication process

■

Must use the same configured AS number

■

Must believe that the source IP address of a received Hello is in that router’s primary connected subnet on that interface

■

K values must match

The wording of the third item in the list bears a little further scrutiny. The primary subnet of an interface is the subnet as implied by the ip address command that does not have the secondary keyword. An EIGRP router looks at the source IP address of a Hello; if the source IP address is a part of that router’s primary subnet of the incoming interface, the Hello passes the IP address check. This logic leaves open some interesting possibilities. For example, if the routers are misconfigured with different subnet masks, the check may still pass. If one router has configured 10.1.2.1/24, and the other has configured 10.1.2.2/23, they could become adjacent, assuming all the other checks pass. While EIGRP supports secondary IP addresses and subnets, EIGRP sources its messages from the address in the primary subnet, and the IP addresses of neighbors must be in the subnet of the primary subnets. The last item in the list mentions K values; K values are constants that define the multipliers used by EIGRP when calculating metrics. The settings can be changed with a router eigrp subcommand metric weights tos k1 k2 k3 k4 k5. The command defaults to a setting of 0 1 0 1 0 0, meaning that only bandwidth and delay are used to calculate the metric. (The examples in this chapter usually use the settings 0 0 0 1 0 0, which removes bandwidth from the calculation and makes the metrics in the examples a little more obvious.) Besides simply checking to see if the right parameters agree, the Hello messages also serve as an EIGRP keepalive. Adjacent routers continue to multicast Hellos based on each interface’s EIGRP hello interval. If a router fails to hear from a neighbor for a number of seconds, defined by the EIGRP Hold Time for that neighbor, all routes through the neighbor are considered to have failed. Example 7-1 shows how a router displays some of the basic information regarding EIGRP operations based on Figure 7-1. The example begins with four routers (R1, R2, S1, and S2) that have only their common LAN interfaces up, just to show the Hello process. By the end of the example, the R2-to-R5 PVC will come up, but the EIGRP adjacency will fail due to a K-value mismatch.

EIGRP Basics and Steady-State Operation

223

Sample Internetwork Used for EIGRP Examples

Figure 7-1

Network 172.31.0.0 14.2/30 S0/0.1

104.4/24 E0/0

14.1/30 S0/0.4

R4

211.202/24

11.1/24

S1

R1

11.201/24

Fa0/0 24.1/30 S0/0.2

VLAN 1 172.31.11.0/24

24.2/30 S0/0.4

11.2/24 11.202/24

105.5/24 E0

Example 7-1

R5

25.5/24 S0

25.2/24 S0/0.5

R2

221.202/24

S2

Fa0/0

Forming EIGRP Adjacencies

! First, a debug is initiated on R1. R1# debug eigrp packet hello EIGRP Packets debugging is on (HELLO) Jan 11 13:27:19.714: EIGRP: Received HELLO on FastEthernet0/0 nbr 172.31.11.201 Jan 11 13:27:19.714:

AS 1, Flags 0x0, Seq 0/0 idbQ 0/0 iidbQ un/rely 0/0 peerQ

un/rely 0/0 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! ! S2’s LAN interface brought up, not shown ! Below, a pair of log messages appear, announcing the new neighbor; this message ! appears due to the default router eigrp subcommand eigrp log-neighbor-changes. Jan 11 13:27:19.995: EIGRP: New peer 172.31.11.202 Jan 11 13:27:19.995: %DUAL-5-NBRCHANGE: IP-EIGRP(0) 1: Neighbor 172.31.11.202 (FastEthernet0/0) is up: new adjacency !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! ! Next, only neighbors who become adjacent—those that pass all the required ! checks for the parameters—are listed. The Hold timer is shown; it starts at ! its maximum, and decrements towards 0, being reset upon the receipt of any EIGRP ! packet from that neighbor. The “H” column on the left states the order in ! which the neighbors became adjacent. R1# show ip eigrp neighbors IP-EIGRP neighbors for process 1 H

Address

Interface

Hold Uptime (sec)

SRTT

RTO

(ms)

Q

Seq

Cnt Num

2

172.31.11.2

Fa0/0

4 00:03:10

1

4500

0

233

1

172.31.11.202

Fa0/0

11 00:04:43

1

4500

0

81

0

172.31.11.201

Fa0/0

14 00:05:11 1927

5000

0

84

! Below, the PVC between R2 and R5 came up, but R5’s K values do not match R2’s. ! Both messages below are log messages, with no debugs enabled on either router. ! Next message on R5 !!!!!!!!!! 03:55:51: %DUAL-5-NBRCHANGE: IP-EIGRP(0) 1: Neighbor 172.31.25.2 (Serial0) is down: K-value mismatch ! Next message on R2 !!!!!!!!!! Jan 11 13:21:45.643: %DUAL-5-NBRCHANGE: IP-EIGRP(0) 1: Neighbor 172.31.25.5 (Serial0/0.5) is down: Interface Goodbye received

224

Chapter 7: EIGRP

Note that when the PVC between R2 and R5 comes up, the message on R5 is pretty obvious, but the message at R2 says nothing about K values. Some later releases of Cisco IOS mistake invalid EIGRP K-value settings as a newer EIGRP message called a Goodbye message. Goodbye messages allow routers to tell each other that they are shutting down in a graceful fashion; be aware that this message may simply be the result of a K-value mismatch. Interestingly, the Hello and Hold time parameters do not need to match for EIGRP neighbor relationships to form. In fact, a router does not use its own timers when monitoring a neighbor relationship—instead, it uses each neighbor’s stated timers, as exchanged in the Hello messages. For example, in Example 7-1, R2 has been configured with Hello and Hold timer settings at 2 and 6 seconds, respectively, with R1 defaulting to 5 and 15 seconds. As R1 monitors its neighbor connection to R2, R1 resets the Hold timer to 6 seconds upon receipt of an EIGRP message. With a hello interval of 2 seconds, R1’s listing for hold time for R2 shows it fluctuating between 6 and 4, assuming no Hellos are lost. Note the show ip eigrp neighbors command on R1 near the end of the example—under normal operation, this value fluctuates between 6 and 4 seconds. The other neighbors default to Hello and Hold time of 5 and 15, so R1’s Hold time in the command output fluctuates between 15 and 10 for these neighbors, assuming no Hellos are lost.

EIGRP Updates Once routers are adjacent, they can exchange routes using EIGRP Update messages. The process follows this general sequence: 1.

Initially, full updates are sent, including all routes except those omitted due to split horizon.

2.

Once all routes have been exchanged, the updates cease.

3.

Future partial updates occur when one or more routes change.

4.

If neighbors fail and recover, or new neighbor adjacencies are formed, full updates are sent.

EIGRP uses the Reliable Transport Protocol (RTP) to send the multicast EIGRP updates. EIGRP sends updates, waiting on a unicast EIGRP ACK message from each recipient. Figure 7-2 shows the general idea over a LAN. Figure 7-2

EIGRP Use of RTP on a LAN 1

3

ACK, to 172.31.11.1

2

Update, to 224.0.0.10

S1

Start RTO Timer

S2

R1 4

3

ACK, to 172.31.11.1

5

Update, to 172.31.11.2

RTO Expired for Neighbor 172.31.11.2 – Resend!

R2

EIGRP Basics and Steady-State Operation

225

RTP allows the Updates to be sent as multicasts. If any neighbors fail to acknowledge receipt of the multicasted update, RTP resends Updates as unicasts just to those neighbors. The steps run as follows, using Figure 7-2 as an example: 1.

The EIGRP sender (R1 in Figure 7-2) starts a Retransmission Timeout (RTO) timer for each neighbor when sending a reliable message like an Update. (Cisco IOS actually calculates a Smoothed Round-Trip Time, or SRTT, to each neighbor, and derives RTO from the SRTT; both values are shown in the show ip eigrp neighbor output. These values vary over time.)

2.

R1 sends the multicast EIGRP Update.

3.

R1 notes from which neighbors it receives an EIGRP ACK for the Update.

4.

RTO expired before router R2 sent its EIGRP ACK.

5.

R1 resends the Update, this time as a unicast, and only to the neighbor(s) that did not reply within the RTO time (R2 in this case).

This process allows efficient multicasting of updates under normal circumstances, and efficient retransmission when ACKs do not arrive in time. EIGRP and RTP use a simple acknowledgement process with a window size of one message. Each Update packet has a sequence number, with the returned ACK message confirming receipt of the message by listing that same sequence number. Example 7-2 shows the location of the sequence number information in both show and debug commands. (In the example, R1 does a no shut on a loopback interface [IP address 172.31.151.1/24], with R1 sending an update advertising the newly-available route.) Example 7-2

Sequence Numbers in EIGRP Updates and ACKs

! First, note the show ip eigrp neighbor output on router R2. The last column ! lists the sequence number last used by that neighbor to send a “reliable” ! packet. So, R2 expects R1’s next reliable EIGRP message to have sequence number ! 225. Also, the RTO calculations are listed for each neighbor. Note ! that the SRTT value is 0 until some reliable packets are exchanged, as SRTT ! is calculated based on actual round-trip time measurements. R2# sh ip eigrp neighbor IP-EIGRP neighbors for process 1 H

Address

Interface

Hold Uptime

SRTT

(sec)

(ms)

RTO

Q

Seq

Cnt Num

2

172.31.11.1

Fa0/0

5 01:14:03

1

200

0

224

1

172.31.11.202

Fa0/0

13 01:15:36

1

200

0

92

0

172.31.11.201

Fa0/0

13 01:16:04

257

1542

0

96

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! ! R1 – R1 – R1 – R1 ! Next, the debug command on R1 enables debug for Update and Ack packets.

continues

226

Chapter 7: EIGRP

Example 7-2

Sequence Numbers in EIGRP Updates and ACKs (Continued)

R1# debug eigrp packet update ack EIGRP Packets debugging is on (UPDATE, ACK) no shutdown,” interface address 172.31.151.1/24. ! Not Shown: R1’s loop0 is “n ! Below, the debug messages show R1’s update, and each of the other three routers’ ! Acks. Note R1’s update has “sequence” 225, and the Acks list that same sequence ! number after the slash. Jan 11 14:43:35.844: EIGRP: Enqueueing UPDATE on FastEthernet0/0 iidbQ un/rely 0/1 serno 207-207 Jan 11 14:43:35.848: EIGRP: Sending UPDATE on FastEthernet0/0 Jan 11 14:43:35.848:

AS 1, Flags 0x0, Seq 225/0 idbQ 0/0 iidbQ un/rely 0/0 serno 207-207

Jan 11 14:43:35.848: EIGRP: Received ACK on FastEthernet0/0 nbr 172.31.11.202 Jan 11 14:43:35.852: 0/1

AS 1, Flags 0x0, Seq 0/225 idbQ 0/0 iidbQ un/rely 0/0 peerQ un/rely

Jan 11 14:43:35.852: EIGRP: Received ACK on FastEthernet0/0 nbr 172.31.11.2 Jan 11 14:43:35.852: 0/1

AS 1, Flags 0x0, Seq 0/225 idbQ 0/0 iidbQ un/rely 0/0 peerQ un/rely

The EIGRP Topology Table EIGRP uses three tables: the neighbor table, the topology table, and the IP routing table. The neighbor table keeps state information regarding neighbors, and is displayed using the show ip eigrp neighbors command. EIGRP Update messages fill the routers’ EIGRP topology tables. Based on the contents of the topology table, each router chooses its best routes and installs these routes in its respective IP routing table. An EIGRP router calculates the metric for each route based on the components of the metric. When a neighboring router advertises a route, the Update includes the metric component values for each route. The router then considers the received metric values, as well its own interfaces settings, to calculate its own metric for each route. The default metric components are cumulative delay, in tens of microseconds, and the constraining bandwidth for the entire route, in bits per second. By setting the correct K values in the metric weights command, EIGRP can also consider link load, reliability, and MTU. Cisco recommends not using those values, in large part due to the fluctuation created by the rapidly changing calculated metrics and repeated routing reconvergence. Figure 7-3 depicts the general logic relating to the metric components in a routing update, showing the units on the bandwidth and delay commands versus the contents of the updates. NOTE A router considers its interface delay settings, as defined with the delay interface subcommand, when calculating EIGRP metrics. The delay command’s units are tens of microseconds, so a delay 1 command sets the interface delay as 10 microseconds.

EIGRP Basics and Steady-State Operation

227

EIGRP Update and Computing the Metric

Figure 7-3

3 Topology Table: Subnet 1 Delay 100 + 1000 = 1100 Bandwidth 1,544 (MTU, Load, Reliability) Calculate My Metric

1 Update: Subnet 1 Delay 100 Bandwidth 10,000 (MTU, Load, Reliability)

Interface Settings: Delay 100 Bandwidth 10,000

R1

2 Calculate RD per Received K Values

4

Interface Settings: Delay 1000 Bandwidth 1544

R2

Update: Subnet 1 Delay 1100 Bandwidth 1,544 (MTU, Load, Reliability)

Metrics: RD = 256 (10,000,000/10,000) + 256 (100) = 281,600 Metric = 256 (10,000,000/1,544) + 256 (1100) = 1,939,631

Because the received update includes the neighbor’s metric components, a router can calculate the advertising neighbor’s metric for a route—called the reported distance (RD). A router can, of course, also calculate its own metric for a particular route, after adding its own interface delay and considering whether it should adjust the value for the constraining bandwidth. For example, consider the four steps outlined in Figure 7-3: 1.

R1 advertises a route, with bandwidth = 10,000 and delay = 100.

2.

R2 calculates the RD for this route per the received K values.

3.

R2 updates its topology table, adding delay 1000 because the interface on which R2 received the update has a delay setting of 1000. It also uses a new bandwidth setting, because the received Update’s bandwidth (10,000) was greater than R2’s incoming interface’s bandwidth (1544).

4.

R2’s update to another neighbor includes the new (cumulative) delay and the new (constraining) bandwidth.

Assuming default K-value settings, the EIGRP formula for the metric calculation is Metric = 256 (107/bandwidth) + 256 (delay) The show ip eigrp topology command lists the RD and the locally computed metric for the entries in the EIGRP topology table. Example 7-3 shows a few details of where the RD and local metric can be seen in show command output. The example is based on Figure 7-1, with all routers and interfaces now working properly. Also, to keep things simple, the delay command has been used to set all links to delay 1 (LANs), delay 2 (WANs), or delay 3 (loopbacks). Also, the metric weights 0 0 0 1 0 0 command was used on each router, taking bandwidth out of the calculation, making the calculated metrics a little more meaningful in the command output.

228

Chapter 7: EIGRP

Example 7-3

EIGRP Topology Table

! First, the numbers in parentheses show this router’s (R1’s) calculated metric, ! then a “/”, then the RD. For example, S1 advertised the route to 211.0/24, with ! R1 calculating S1’s metric (the RD) as 768. Delay 3 was set on S1’s loopback ! (where 211.0/24 resides), so its metric was 3*256=768. R1’s metric adds delay 1, ! for a metric of 4*256=1024. R1# show ip eigrp topology IP-EIGRP Topology Table for AS(1)/ID(172.31.16.1) Codes: P - Passive, A - Active, U - Update, Q - Query, R - Reply, r - reply Status, s - sia Status P 172.31.151.0/24, 1 successors, FD is 768 via Connected, Loopback1 P 172.31.211.0/24, 1 successors, FD is 1024 via 172.31.11.201 (1024/768), FastEthernet0/0 P 172.31.24.0/30, 1 successors, FD is 768 via 172.31.11.2 (768/512), FastEthernet0/0 via 172.31.14.2 (1024/512), Serial0/0.4 ! Lines omitted for brevity ! Below, the metric in the IP routing table entries match the first number in ! the parentheses, as well as the number listed as “FD is…” in the output above. R1# show ip route ! omitted legend for brevity 172.31.0.0/16 is variably subnetted, 9 subnets, 2 masks D

172.31.211.0/24 [90/1024] via 172.31.11.201, 00:29:42, FastEthernet0/0

D

172.31.24.0/30 [90/768] via 172.31.11.2, 00:29:44, FastEthernet0/0

! Lines omitted for brevity

The show ip eigrp topology command lists a few additional very important concepts and terms related to how EIGRP chooses between multiple possible routes to the same prefix. First, the term feasible distance (FD) refers to this router’s best calculated metric among all possible routes to reach a particular prefix. The FD is listed as “FD is x” in the command output. The route that has this best FD is called the successor route, and is installed in the routing table. The successor route’s metric is by definition called the feasible distance, so that metric is what shows up in the routes shown with the show ip route command. These additional terms all relate to how EIGRP processes convergence events, which is explained next.

EIGRP Convergence Once all the EIGRP routers have learned all the routes in the network, and placed the best routes (the successor routes) in their IP routing tables, their EIGRP processes simply continue to send Hellos, expect to receive Hellos, and look for any changes to the network. When those changes do occur, EIGRP must converge to use the best available routes. This section covers the three major

EIGRP Convergence

229

components of EIGRP convergence: input events, local computation (which includes looking for feasible successors), and using active querying to find alternative routes. Table 7-3 lists several of the key EIGRP terms related to convergence. Following the table, the text jumps right into what EIGRP does when a topology or metric change occurs. EIGRP Features Related to Convergence

Table 7-3

EIGRP Convergence Function

Description

Reported distance (RD)

The metric (distance) of a route as reported by a neighboring router

Feasible distance (FD)

The metric value for the lowest-metric path to reach a particular subnet

Feasibility condition

When multiple routes to reach one subnet exist, the case in which one route’s RD is lower than the FD

Successor route

The route to each destination prefix for which the metric is the lowest metric

Feasible successor (FS)

A route that is not a successor route but meets the feasibility condition; can be used when the successor route fails, without causing loops

Input event

Any occurrence that could change a router’s EIGRP topology table

Local computation

An EIGRP router’s reaction to an input event, leading to the use of a feasible successor or going active on a route

Input Events and Local Computation An EIGRP router needs to react when an input event occurs. The obvious input events are when a router learns of new prefixes via newly received routing updates, when an interface fails, or when a neighbor fails. Because EIGRP sends updates only as a result of changed or new topology information, a router must consider the update and decide if any of its routes have changed. When an input event implies that a route has failed, the router performs local computation, a fancy term for a process that can be boiled down to relatively simple logic. In short, the result of local computation is that the router either is able to choose a replacement route locally, without having to ask any neighbors, or is required to ask neighbors for help. Simply put, for a failed route, local computation does the following: ■

If FS routes exist, install the lowest-metric FS route into the routing table, and send Updates to neighbors to notify them of the new route.

■

If no FS route exists, actively query neighbors for a new route.

To be an FS route, a route must meet the feasibility condition, defined as follows: The RD must be lower than this router’s current FD for the route.

230

Chapter 7: EIGRP

The local computation is best understood by looking at an example. Figure 7-4 shows the same network as in Figure 7-1, but with delay values shown. Example 7-4 begins with R4 using a successor route to 172.31.211.0/24, through R1. R4 also has an FS route to 172.31.211.0/24 through R2. The example shows what happens when the PVC from R1 to R4 fails, and R4’s neighbor relationship with R1 fails, causing R4 to perform local computation and start using its FS route through R2. Network Used for EIGRP Convergence Examples

Figure 7-4

14.2/30 S0/0.1

104.4/24 E0/0

Delay: 3

Delay: 3 151.1/24

14.1/30 S0/0.4

R4

Delay: 3 211.202/24

S1

R1 Delay: 2 Fa0/0 11.1/24 24.1/30 S0/0.2

11.201/24

Delay: 3

VLAN 1 24.2/30 S0/0.4

172.31.11.0/24 11.2/24

105.5/24 E0

R5

25.5/24 S0

Delay: 3

25.2/24 S0/0.5

R2

Delay: 4

Delay: 3 221.202/24

11.202/24

S2

NOTE The routers have disabled the use of bandwidth in the EIGRP metric calculation, so all metrics in Example 7-4 are multiples of 256.

Example 7-4

Local Computation: R1-R4 Link Fails; R4 Finds an FS to 172.31.211.0/24 Through R2

! First, the current successor route on R4 points out S0/0.1, to R1, metric 2048. R4# show ip route ! lines omitted for brevity 172.31.0.0/16 is variably subnetted, 9 subnets, 2 masks D

172.31.211.0/24 [90/2048] via 172.31.14.1, 00:01:46, Serial0/0.1

! Below, the FD is listed as 2048 as well. The topology entry for the successor ! has the same 2048 metric listed as the first number in parentheses; the second ! number is the RD on R1 (1280). The second topology entry for this route lists ! metric 2560, RD 1792; with RD in the second route being less than the FD, this ! second route meets the feasibility condition, making it an FS route. R4# show ip eigrp topology IP-EIGRP Topology Table for AS(1)/ID(172.31.104.4) Codes: P - Passive, A - Active, U - Update, Q - Query, R - Reply, r - reply Status, s - sia Status ! lines omitted for brevity P 172.31.211.0/24, 1 successors, FD is 2048 via 172.31.14.1 (2048/1280), Serial0/0.1 via 172.31.24.2 (2560/1792), Serial0/0.2 ! Next, R4 loses Neighbor R1, with EIGRP Finite State Machine (FSM) debug on. R4# debug eigrp fsm

EIGRP Convergence

Example 7-4

231

Local Computation: R1-R4 Link Fails; R4 Finds an FS to 172.31.211.0/24 Through R2 (Continued)

EIGRP FSM Events/Actions debugging is on Jan 12 07:17:42.391: %DUAL-5-NBRCHANGE: IP-EIGRP(0) 1: Neighbor 172.31.14.1 (Serial0/0.1) is down: holding time expired ! Below, debug messages have been edited to only show messages relating to ! the route to 172.31.211.0/24. R4 looks for an FS, finds it, replaces the old ! successor with the FS, and sends updates telling neighbors about the new route. Jan 12 07:17:42.399: DUAL: Destination 172.31.211.0/24 Jan 12 07:17:42.399: DUAL: Find FS for dest 172.31.211.0/24. FD is 2048, RD is 2048 Jan 12 07:17:42.399: DUAL: 172.31.14.1 metric 4294967295/4294967295 Jan 12 07:17:42.399: DUAL: 172.31.24.2 metric 2560/1792 found Dmin is 2560 Jan 12 07:17:42.399: DUAL: Removing dest 172.31.211.0/24, nexthop 172.31.14.1 Jan 12 07:17:42.403: DUAL: RT installed 172.31.211.0/24 via 172.31.24.2 Jan 12 07:17:42.403: DUAL: Send update about 172.31.211.0/24.

Reason: metric chg

Jan 12 07:17:42.403: DUAL: Send update about 172.31.211.0/24.

Reason: new if

! Finally, note that the FD is unchanged; the FD is never raised until the route ! has been actively queried. The new route info has been put in the routing table. R4# show ip eigrp topology ! lines omitted for brevity P 172.31.211.0/24, 1 successors, FD is 2048 via 172.31.24.2 (2560/1792), Serial0/0.2 R4# show ip route ! Lines omitted for brevity D

172.31.211.0/24 [90/2560] via 172.31.24.2, 00:00:25, Serial0/0.2

Going Active on a Route The second branch in the local computation logic causes the EIGRP router to ask its neighbors about their current best route to a subnet, hoping to find an available, loop-free alternative route to that subnet. When no FS route is found, the EIGRP router goes active for the route. Going active is jargon for the process of changing a route’s status to active. Once the router is active, EIGRP multicasts Query messages to its neighbors, asking the neighbors if they have a valid route to the subnet. The neighbors should unicast EIGRP Reply packets back to the original router, stating whether or not they have a current loop-free route with which to reach that prefix. Once a router receives Reply messages from all the neighbors to which it sent Queries, the router updates its topology table with all the new information learned in the Reply messages, recomputes metrics for any known routes, and chooses a new successor. Of course, if no routes to that subnet are found, this router simply does not add a route to the routing table. NOTE The EIGRP term “active” refers to a route for which a router is currently using the Query process to find a loop-free alternative route. Conversely, a route is in passive state when it is not in an active state.

232

Chapter 7: EIGRP

The neighboring routers view any received Query messages as an input event. Each neighbor router’s behavior when receiving a Query can be summarized as follows: 1.

If the router does not have an entry in its topology table for that subnet, it sends an EIGRP Reply packet stating that it has no route.

2.

If the router’s successor for that subnet is unchanged, or an FS is found, the neighbor sends back an EIGRP Reply message with the details of the route.

3.

If the conditions in step 1 or 2 do not exist, the router itself goes active, and withholds its EIGRP response to the original Query, until all of its neighbors respond.

Note that the logic in the third step can result in a route for which the Active Querying process never completes. Routes that stay in active state too long are considered to be stuck-in-active routes. The related concepts are covered in the next section. Example 7-5 shows an example of the Query process. The example is again based on Figure 7-4, with R4 again losing its neighbor relationship with R1. In this case, R4’s local computation will not find an FS for its failed route to 172.31.151.0/24, so it must go active. Example 7-5

R1-R4 Link Fails; R4 Actively Queries for 172.31.151.0/24

! First, the show ip eigrp topology command only lists the successor route, and no ! FS routes. This command does not list non-FS routes. R4# show ip eigrp topo ! Lines omitted for brevity P 172.31.151.0/24, 1 successors, FD is 1536 via 172.31.14.1 (1536/768), Serial0/0.1 ! Below, the show ip eigrp topology all-links command includes non-FS routes, ! in this case including the non-FS route to 151.0/24 through R2. Note that this ! alternate non-FS route’s RD is 1792, which is more than the FD of 1536. R4# show ip eigrp topology all-links ! Lines omitted for brevity P 172.31.151.0/24, 1 successors, FD is 1536, serno 175 via 172.31.14.1 (1536/768), Serial0/0.1 via 172.31.24.2 (2560/1792), Serial0/0.2 ! Next, the FSM debug is again enabled, and R4 loses neighbor R1. R4# debug eigrp fsm Jan 12 07:16:04.099: %DUAL-5-NBRCHANGE: IP-EIGRP(0) 1: Neighbor 172.31.14.1 (Serial0/0.1) is down: holding time expired !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! ! Below, R4 looks for an FS for route 172.31.151.0/24, and does not find one— ! so it enters active state. R4 sends a query to its one remaining neighbor (R2), ! and keeps track of the number of outstanding Queries (1). Upon receiving the ! Reply from R2, it can update its topology table, and repeat local computation, ! and use the now-best route through R2. Jan 12 07:17:42.391: %DUAL-5-NBRCHANGE: IP-EIGRP(0) 1: Neighbor 172.31.14.1 (Serial0/0.1) is down: holding time expired

EIGRP Convergence

Example 7-5

233

R1-R4 Link Fails; R4 Actively Queries for 172.31.151.0/24 (Continued)

Jan 12 07:17:42.391: DUAL: linkdown: start - 172.31.14.1 via Serial0/0.1 Jan 12 07:17:42.391: DUAL: Destination 172.31.151.0/24 Jan 12 07:17:42.391: DUAL: Find FS for dest 172.31.151.0/24. FD is 1536, RD is 1536 Jan 12 07:17:42.395: DUAL:

172.31.14.1 metric 4294967295/4294967295

Jan 12 07:17:42.395: DUAL:

172.31.24.2 metric 2560/1792 not found Dmin is 2560

Jan 12 07:17:42.395: DUAL: Dest 172.31.151.0/24 entering active state. Jan 12 07:17:42.395: DUAL: Set reply-status table. Count is 1. Jan 12 07:17:42.395: DUAL: Not doing split horizon Jan 12 07:17:42.459: DUAL: rcvreply: 172.31.151.0/24 via 172.31.24.2 metric 2560/1792 Jan 12 07:17:42.459: DUAL: reply count is 1 Jan 12 07:17:42.459: DUAL: Clearing handle 0, count now 0 Jan 12 07:17:42.463: DUAL: Freeing reply status table Jan 12 07:17:42.463: DUAL: Find FS for dest 172.31.151.0/24. FD is 4294967295, RD is 4294967295 found Jan 12 07:17:42.463: DUAL: Removing dest 172.31.151.0/24, nexthop 172.31.14.1 Jan 12 07:17:42.463: DUAL: RT installed 172.31.151.0/24 via 172.31.24.2 Jan 12 07:17:42.467: DUAL: Send update about 172.31.151.0/24.

Reason: metric chg

Jan 12 07:17:42.467: DUAL: Send update about 172.31.151.0/24.

Reason: new if

! Next, note that because R4 actively queried for the route, the FD could change. R4# show ip eigrp topo IP-EIGRP Topology Table for AS(1)/ID(172.31.104.4) Codes: P - Passive, A - Active, U - Update, Q - Query, R - Reply, r - reply Status, s - sia Status P 172.31.151.0/24, 1 successors, FD is 2560 via 172.31.24.2 (2560/1792), Serial0/0.2

Of particular note in this example, look for the debug message starting with “Dual: rcvreply:” (highlighted). This message means that the router received an EIGRP Reply message, in this case from R2. The message includes R2’s valid routing information for 172.31.151.0/24. Also note that the FD was recomputed, whereas it was not in Example 7-4 when an FS route was found. NOTE Query messages use reliable transmission via RTP and are multicasts; Reply messages are reliable and are unicasts. Both are acknowledged using Ack messages. NOTE The EIGRP term Diffusing Update Algorithm (DUAL) refers to the totality of the logic used by EIGRP to calculate new routes. The term is based on the logic used as Query messages go outward from a router, with the outward movement stopped when routers Reply.

Stuck-in-Active Any router in active state for a route must wait for a Reply to each of its Query messages. It is possible for a router to wait several minutes for all the replies, because neighboring routers might also need to go active, and then their neighbors might need to go active, and so on—each withholding its Reply message until it in turn receives all of its Reply messages. In normal

234

Chapter 7: EIGRP

operation, the process should complete; to handle exception cases, EIGRP includes a timer called the Active timer, which limits the amount of time in which a route can stay active. If the Active timer expires before a router receives all of its Reply messages, the router places the route in a stuck-in-active state. The router also brings down any neighbors from which no corresponding Reply was received, thinking that any neighbors that did not send a Reply are having problems. In some conditions—large, redundant networks, flapping interfaces, or networks with lots of packet loss, to name a few—neighbors might be working fine, but their Reply messages may not complete within the Active timer. To avoid the downside of having the route become stuck-in-active, and losing all routes through a possibly still-working neighbor, you can disable the Active timer by using the timers active-time disabled subcommand under router eigrp. Limiting Query Scope Although disabling the Active timer can prevent stuck-in-active routes, a better solution to the prolonged wait for Reply messages is to limit the scope of Query messages. By reducing the number of neighbors that receive the messages, and by limiting the number of hops away the queries flow, you can greatly reduce the time required to receive all Reply messages. Two methods can be used to limit query scope. The first is route summarization. When a Query reaches a router that has a summarized route, but not the specific route in the query, the router immediately replies that it does not have that route. For instance, a router with the route 172.31.0.0/16 in its topology table, upon receiving a query for 172.31.151.0/24, immediately sends a Reply, stating it does not have a route to 172.31.151.0/24. With well-designed route summarization, EIGRP queries can be limited to a few hops. (Chapter 9 covers route summarization details.) The use of EIGRP stub routers also limits the query scope. Stub routers, by definition, should not be used as transit routers for traffic. In Figure 7-4, R5 would be a classic candidate to be a stub router. Also, if R4 should not be used to forward traffic from R1 over to R2, or vice versa, R4 could be a stub as well. In either case, non-stub routers do not send Query messages to the stub routers, knowing that the stub routers should not be transit routers. (Stub router configuration is covered in the next section.)

EIGRP Configuration This section explains the majority of the options for EIGRP configuration. The “Foundation Summary” section includes the full syntax of the commands, along with some comments, in Table 7-6.

EIGRP Configuration Example Example 7-6 lists the configuration for R1, R2, R4, and R5 from Figure 7-4. The routers were configured based on the following design goals: ■

Enable EIGRP on all interfaces.

EIGRP Configuration

235

■

Configure K values to ignore bandwidth.

■

Configure R5 as an EIGRP stub router.

■

Ensure that R2’s LAN interface uses a Hello and Hold time of 2 and 6, respectively.

■

Configure R4 to allow 75 percent of interface bandwidth for EIGRP updates.

■

Advertise R4’s LAN subnet, but do not attempt to send or receive EIGRP updates on the LAN.

Example 7-6

Basic EIGRP Configuration on R1, R2, R4, and R5

! Below, R1 EIGRP-related configuration ! The default metric weights are “0 1 0 1 0 0”. router e igrp 1 n e t w o r k 1 72 . 3 1 . 0 . 0 metr ic weights 0 0 0 1 0 0 ! R2 EIGRP-related configuration ! Note the commands used to change the Hello and Hold Time values per interface. ! R2’s Hellos advertise the timer values, and other routers on the LAN use these ! values on their neighbor relationship with R2. Also below, note the use of the ! inverse mask to match a subset of interfaces on a single network command. int erface FastEther net0/0 ip hello-inte rval eigrp 1 2 ip hold-time eigrp 1 6 ! router eigrp 1 network 10.0.0.0 n e t w o r k 1 7 2 . 3 1 . 1 1 .2 0 . 0 . 0 . 0 n e t w o r k 1 7 2 . 3 1 .2 4 . 0 0 . 0 . 1 . 2 5 5 m e t r i c w e i gh t s 0 0 0 1 0 0 ! R4 EIGRP-related configuration ! Below, the percentage of the interface bandwidth used for EIGRP is changed. The ! value can go over 100% to allow for cases in which the bandwidth has ! been artificially lowered to impact the EIGRP metric. Also note that R4 makes ! its e0/0 interface passive, meaning no routes learned or advertised on E0/0. i n t e r f a c e S e r i a l 0 / 0 . 1 p o in t - t o - p o i n t bandwidth 64 ip bandwi dth-percent eigrp 1 150 ! router e igrp 1 p a s s i v e - i nt e r f a c e E t h e r n e t 0 / 0 n e tw o r k 1 7 2 . 3 1 . 0 . 0 metric weights 0 0 0 1 0 0 ! R5 EIGRP-related configuration ! Below, note R5’s configuration as a stub area.

continues

236

Chapter 7: EIGRP

Example 7-6

Basic EIGRP Configuration on R1, R2, R4, and R5 (Continued)

router e igrp 1 n e t w o r k 1 72 . 3 1 . 0 . 0 metr ic weights 0 0 0 1 0 0 e i g r p s t u b c o nn e c t e d s u m m a r y

EIGRP allows for better control of the three functions enabled on an interface by the EIGRP network command. (The three functions are advertising the connected subnet, sending routing updates, and receiving routing updates.) Like OSPF, the EIGRP network command supports configuration of an optional wildcard mask (as seen on R4 in Example 7-6), allowing each interface to be matched individually—and making it simple to enable EIGRP on a subset of interfaces. Also, a LAN subnet might have a single router attached to it, so there is no need to attempt to send or receive updates on those interfaces. By enabling EIGRP on the interface with a network command, and then configuring the passive-interface command, you can stop the router from sending Hellos. If a router does not send Hellos, it forms no neighbor adjacencies, and it then neither sends nor receives updates on that LAN. Example 7-6 also shows R5 configured as an EIGRP stub router. R5 announces itself as a stub router via its EIGRP Hellos. As a result, R2 will not send Query messages to R5, limiting the scope of Query messages. The eigrp stub command has several options, with the default options (connected and summary) shown on the last line of Example 7-6. (Note that the eigrp stub command was typed, and IOS added the connected and summary options in the configuration.) Table 7-4 lists the eigrp stub command options, and explains some of the logic behind using them. Table 7-4

Options on the eigrp stub Command Option

This Router Is Allowed To…

connected

Advertise connected routes, but only for interfaces matched with a network command.

summary

Advertise auto-summarized or statically configured summary routes.

static

Advertise static routes, assuming the redistribute static command is configured.

redistributed

Advertise redistributed routes, assuming redistribution is configured.

receive-only

Not advertise any routes. This option cannot be used with any other option.

Note that the stub option still requires the stub router to form neighbor relationships, even in receive-only mode. The stub router simply performs less work and reduces the query scope. Example 7-6 also shows the EIGRP hello interval and hold time being set. These parameters can be set per interface using the interface subcommands ip hello-interval eigrp asn seconds and

EIGRP Configuration

237

ip hold-time eigrp asn seconds, respectively. The default EIGRP hello interval defaults to 5 seconds on most interfaces, with NBMA interfaces whose bandwidth is T1 or slower using a hello interval of 60 seconds. The hold time defaults to 15 and 180 seconds, respectively—three times the default hello interval. However, if you change the hello interval, the hold time default does not automatically change to three times the new hello interval; instead, it remains at 15 or 180 seconds.

EIGRP Load Balancing EIGRP allows for up to six equal-metric routes to be installed into the IP routing table at the same time. However, because of the complex EIGRP metric calculation, metrics may often be close to each other, but not exactly equal. To allow for metrics that are somewhat close in value to be considered equal, and added to the IP routing table, you can use the variance multiplier command. The multiplier defines a value that is multiplied by the lowest metric (in other words, the FD, which is the metric of the successor route). If any other routes have a better metric than that product of variance * FD, those other routes are considered equal, and added to the routing table. NOTE EIGRP allows only FS routes to be considered for addition as a result of using the variance command. Otherwise, routing loops could occur.

Once the multiple routes for the same destination are in the routing table, EIGRP allows several options for balancing traffic across the routes. Table 7-5 summarizes the commands that impact how load balancing is done with EIGRP, plus the other commands related to installing multiple EIGRP routes into the same subnet. Note that these commands are all subcommands under router eigrp. Table 7-5

EIGRP Route Load-Balancing Commands Router EIGRP Subcommand

Meaning

variance

Any FS route whose metric is less than the variance value multiplied by the FD is added to the routing table (within the restrictions of the maximumpaths command).

maximum-paths {1..6}

The maximum number of routes to the same destination allowed in the routing table. Defaults to 4.

traffic-share balanced

The router balances across the routes, giving more packets to lower-metric routes.

traffic-share min

Although multiple routes are installed, sends traffic using only the lowestmetric route.

traffic-share min across-interfaces

If more than 1 route has the same metric, the router chooses routes with different outgoing interfaces, for better balancing.

No traffic-share command configured

Balances evenly across routes, ignoring EIGRP metrics.

238

Chapter 7: EIGRP

EIGRP Authentication EIGRP authentication, much like OSPF authentication, requires the creation of keys and requires authentication to be enabled on a per-interface basis. The keys are used as the secret (private) key used in an MD5 calculation. (EIGRP does not support clear-text authentication.) Multiple keys are allowed and are grouped together using a construct called a key chain. A key chain is simply a set of related keys, each of which has a different number and may be restricted to a time period. By allowing multiple related keys in a key chain, with each key valid during specified time periods, the engineer can easily plan for migration to new keys in the future. (NTP is recommended when keys are restricted by time ranges, because the local times on the routers must be synchronized for this feature to work correctly.) Cisco IOS enables the EIGRP authentication process on a per-interface basis using the command ip authentication mode eigrp asn md5, and refers to the key chain that holds the keys with the ip authentication key-chain eigrp asn key_name interface subcommand. The router looks in the key chain and selects the key(s) valid at that particular time. Example 7-7 shows the EIGRP authentication configuration for R1, R2, and R4, and includes a few additional comments. The network in Figure 7-1 is the basis for this example. Example 7-7

EIGRP Authentication (R1, R2, and R4)

! First, R1 Config ! Chain “carkeys” will be used on R1’s LAN. R1 will use key “fred” for ! about a month, and then start using “wilma.” key chain carkeys key 1 key-string fred accept-lifetime 08:00:00 Jun 11 2007 08:00:00 Jul 11 2007 send-lifetime 08:00:00 Jun 11 2007 08:00:00 Jul 11 2007 key 2 key-string wilma accept-lifetime 08:00:00 Jul 10 2007 08:00:00 Aug 11 2007 send-lifetime 08:00:00 Jul 10 2007 08:00:00 Aug 11 2007 ! Next, key chain “anothersetofkeys” defines the key to be ! used with R4. key chain anothersetofkeys key 1 key-string barney ! Next, R1’s interface subcommands are shown. ! The key chain is referenced ! using the ip eigrp 1 authentication command. interface FastEthernet0/0 ip address 172.31.11.1 255.255.255.0 ip authentication mode eigrp 1 md5 ip authentication key-chain eigrp 1 carkeys ! Below, R1 enables EIGRP authentication on

EIGRP Configuration

Example 7-7

239

EIGRP Authentication (R1, R2, and R4) (Continued)

! the subinterface connecting to R4. interface Serial0/0.4 point-to-point ip address 172.31.14.1 255.255.255.252 ip authentication mode eigrp 1 md5 ip authentication key-chain eigrp 1 anothersetofkeys ! R2 Config – R2 Config – R2 Config ! Next, on R2, the key chain name (housekeys) differs with ! R1’s key chain name (carkeys), but ! the key string “fred” is the same. key chain housekeys key 1 key-string fred interface FastEthernet0/0 ip address 172.31.11.2 255.255.255.0 ip authentication mode eigrp 1 md5 ip authentication key-chain eigrp 1 housekeys ! R4 Config – R4 Config – R4 Config ! Next, R4 enables EIGRP authentication on its subinterface connecting to R1. key chain boatkeys key 1 key-string barney ! interface Serial0/0.1 point-to-point ip address 172.31.14.2 255.255.255.252 ip authentication mode eigrp 1 md5 ip authentication key-chain eigrp 1 boatkeys

Although the comments in Example 7-7 explain the more important details, one other point needs to be made regarding the key lifetimes. The configuration shows that two of the keys’ lifetimes overlap by a day. On that day, EIGRP would use the key with the lowest key number. By using such logic, you could start by configuring one key. Later, you could then add a second key on all the routers, with overlapping time periods, but still use the original key. Finally, you could either let the first key expire or delete the first key, allowing for easy key migration.

EIGRP Automatic Summarization EIGRP defaults to use automatic summarization, or autosummarization. Autosummarization can be disabled with the no auto-summary command under router eigrp process. Unless you particularly want a router to autosummarize using EIGRP, you should configure the no autosummary command to disable this feature. (Note that EIGRP autosummarization works the same in concept as autosummarization with RIP.)

240

Chapter 7: EIGRP

EIGRP Split Horizon EIGRP bounds its updates using split-horizon logic. Split horizon can be disabled on a perinterface basis by using the no ip split-horizon eigrp asn interface subcommand. Most interface types enable split horizon by default, with the notable exception of a physical serial interface configured for Frame Relay.

EIGRP Route Filtering Outbound and inbound EIGRP updates can be filtered at any interface, or for the entire EIGRP process. To filter the routes, the distribute-list command is used under router eigrp asn, referencing an IP ACL. The generic command, when creating an EIGRP distribution list that uses an ACL, is distribute-list {access-list-number | name} {i in | out} [interface-type interfacenumber]

Example 7-8 shows an inbound distribution list on router R2 (in the example in Figure 7-1), filtering routes in the 172.31.196.0/22 range. For this example, R2 now receives several /24 and /30 routes from S2, using EIGRP. The routes are in the range of 172.31.192.0/21, and the goal is to filter the upper half of that numeric range. Example 7-8

EIGRP Distribution List

! The example begins with a list of the routes that should be filtered. ! Note that the longer-prefixes option below makes the command ! list all routes in the range. ! The highlighted lines are the ones that will be filtered. R2# show ip route 172.31.192.0 255.255.248.0 longer-prefixes ! Lines omitted for brevity; in this case, the legend was deleted 172.31.0.0/16 is variably subnetted, 24 subnets, 3 masks D

172.31.195.0/30 [120/1] via 172.31.11.202, 00:00:18, FastEthernet0/0

D

172.31.194.0/24 [120/1] via 172.31.11.202, 00:00:18, FastEthernet0/0

D

172.31.196.4/30 [120/1] via 172.31.11.202, 00:00:18, FastEthernet0/0

D

172.31.195.4/30 [120/1] via 172.31.11.202, 00:00:18, FastEthernet0/0

D

172.31.197.0/24 [120/1] via 172.31.11.202, 00:00:19, FastEthernet0/0

D

172.31.196.0/30 [120/1] via 172.31.11.202, 00:00:19, FastEthernet0/0

D

172.31.195.8/30 [120/1] via 172.31.11.202, 00:00:19, FastEthernet0/0

! R2’s Configuration follows. access-list 2 denies all subnets in the ! 172.31.196.0/22 range, which is the set of subnets that needs to be filtered. ! The distribute-list 2 in FastEthernet0/0 command tells EIGRP to filter inbound ! EIGRP updates that come in fa0/0. router eigrp 1 network 10.0.0.0 network 172.31.0.0 distribute-list 2 in FastEthernet0/0 ! access-list 2 deny

172.31.196.0 0.0.3.255

access-list 2 permit any

EIGRP Configuration

Example 7-8

241

EIGRP Distribution List (Continued)

! Below, the results show three less subnets in the larger 172.31.192.0/21 range. R2# show ip route 172.31.192.0 255.255.248.0 longer-prefixes ! Lines omitted for brevity; in this case, the legend was deleted 172.31.0.0/16 is variably subnetted, 21 subnets, 3 masks D

172.31.195.0/30 [90/1] via 172.31.11.202, 00:00:22, FastEthernet0/0

D

172.31.194.0/24 [90/1] via 172.31.11.202, 00:00:22, FastEthernet0/0

D

172.31.195.4/30 [90/1] via 172.31.11.202, 00:00:22, FastEthernet0/0

D

172.31.195.8/30 [90/1] via 172.31.11.202, 00:00:22, FastEthernet0/0

An EIGRP distribute list might refer to a prefix list instead of an ACL to match routes. Prefix lists are designed to match a range of subnets, as well as a range of subnet masks associated with the subnets. The distribute list must still define the direction of the updates to be examined (in or out), and optionally an interface. Chapter 9 includes a more complete discussion of the syntax and formatting of prefix lists; this chapter focuses on how to call and use a prefix list for EIGRP route filtering. To reference a prefix list, use the following router eigrp asn subcommand: distribute-list {p prefix list-name} {i in | out} [interface-type interface-number]

Example 7-9 shows the execution of this syntax, with the prefix list denying all /30 routes from the range 172.31.192.0/21. The prefix list permits all other subnets. Example 7-9

EIGRP Prefix Lists

! The example begins with a list of the routes that should be filtered. ! Note that the longer-prefixes option below makes the ! command list all routes in the range. ! The highlighted lines are the ones that will be filtered. R2# show ip route 172.31.192.0 255.255.248.0 longer-prefixes ! Lines omitted for brevity; in this case, the legend was deleted 172.31.0.0/16 is variably subnetted, 24 subnets, 3 masks D

172.31.195.0/30 [90/1] via 172.31.11.202, 00:00:18, FastEthernet0/0

D

172.31.194.0/24 [90/1] via 172.31.11.202, 00:00:18, FastEthernet0/0

D

172.31.196.4/30 [90/1] via 172.31.11.202, 00:00:18, FastEthernet0/0

D

172.31.195.4/30 [90/1] via 172.31.11.202, 00:00:18, FastEthernet0/0

D

172.31.197.0/24 [90/1] via 172.31.11.202, 00:00:19, FastEthernet0/0

D

172.31.196.0/30 [90/1] via 172.31.11.202, 00:00:19, FastEthernet0/0

D

172.31.195.8/30 [90/1] via 172.31.11.202, 00:00:19, FastEthernet0/0

! R2’s configuration follows. The “wo2” prefix list limits the mask range to ! only /30 with the “ge 30 le 30” parameters. It matches any subnets between ! 172.31.192.0 and 172.31.199.255. ! Note that the prefix-list commands are global commands. router eigrp 1 network 10.0.0.0 network 172.31.0.0 distribute-list prefix wo2 in FastEthernet0/0

continues

242

Chapter 7: EIGRP

Example 7-9

EIGRP Prefix Lists (Continued)

! ip prefix-list wo2 seq 5 deny 172.31.192.0/21 ge 30 le 30 ip prefix-list wo2 seq 10 permit 0.0.0.0/0 le 32 ! Below, note the absence of /30 routes in the specified range, and the presence ! of the two /24 routes seen at the beginning of Example 8-8. R2# show ip route 172.31.192.0 255.255.248.0 longer-prefixes ! Lines omitted for brevity; in this case, the legend was deleted 172.31.0.0/16 is variably subnetted, 19 subnets, 3 masks D

172.31.194.0/24 [90/1] via 172.31.11.202, 00:00:23, FastEthernet0/0

D

172.31.197.0/24 [90/1] via 172.31.11.202, 00:00:23, FastEthernet0/0

One key concept is worth noting before we move on: With EIGRP filtering, an incoming filter prevents topology information from entering the EIGRP topology table. That is, inbound filters do not affect the routing table directly, but because they keep routing information from the topology table, they have the same effect.

EIGRP Offset Lists EIGRP offset lists allow EIGRP to add to a route’s metric, either before sending an update, or for routes received in an update. The offset list refers to an ACL (standard, extended, or named) to match the routes; any matched routes have the specified offset, or extra metric, added to their metrics. Any routes not matched by the offset list are unchanged. The offset list also specifies which routing updates to examine by specifying a direction (in or out) and, optionally, an interface. If the interface is omitted from the command, all updates for the defined direction will be examined.

Offset lists are much more applicable to RIP (version 1 or 2) than EIGRP because RIP has such a limited metric range. With EIGRP, because of the metric’s complexity, it is doubtful that you would manipulate EIGRP metrics this way. Because several other filtering methods and ways to influence EIGRP metrics are available, offset lists see limited use in EIGRP and are therefore not covered in more detail in this chapter.

EIGRP Configuration

243

Clearing the IP Routing Table The clear ip route * command clears the IP routing table. However, because EIGRP keeps all possible routes in its topology table, a clear ip route * command does not cause EIGRP to send any messages or learn any new topology information; the router simply refills the IP routing table with the best routes from the existing topology table. The clear ip eigrp neighbor command clears all neighbor relationships, which clears the entire topology table on the router. The neighbors then come back up, send new updates, and repopulate the topology and routing tables. The clear command also allows for clearing all neighbors that are reachable out an interface, or based on the neighbor’s IP address. The generic syntax is clear ip eigrp neighbors [ip-address | interface-type interface-number]

244

Chapter 7: EIGRP

Foundation Summary This section lists additional details and facts to round out the coverage of the topics in this chapter. Unlike most of the Cisco Press Exam Certification Guides, this “Foundation Summary” does not repeat information presented in the “Foundation Topics” section of the chapter. Please take the time to read and study the details in the “Foundation Topics” section of the chapter, as well as review items noted with a Key Topic icon. Table 7-6 lists some of the most popular Cisco IOS commands related to the topics in this chapter. Also refer to Table 7-4 for a few additional commands related to load balancing. Table 7-6

Command Reference for Chapter 7 Command

Command Mode and Description

router eigrp as-number

Global config; puts user in EIGRP configuration mode for that AS

network ip-address [wildcard-mask]

EIGRP config mode; defines matching parameters, compared to interface IP addresses, to pick interfaces on which to enable EIGRP

ip split-horizon eigrp asn

Interface subcommand; enables or disables split horizon

passive-interface [default] {interfacetype interface-number}

EIGRP config mode; causes EIGRP to stop sending Hellos on the specified interface, and thereby to also stop receiving and/or sending updates

ip hello-interval eigrp asn seconds

Interface subcommand; sets the interval for periodic Hellos sent by this interface

ip hold-time eigrp asn seconds

Interface subcommand; sets the countdown timer to be used by a router’s neighbor when monitoring for incoming EIGRP messages from this interface

auto-summary

EIGRP config mode; enables automatic summarization at classful network boundaries

metric weights tos k1 k2 k3 k4 k5

EIGRP config mode; defines the per-ToS K values to be used in EIGRP metric calculations

ip bandwidth-percent eigrp asn percent

Interface subcommand; defines the maximum percentage of interface bandwidth to be used for EIGRP messages

Foundation Summary

Table 7-6

245

Command Reference for Chapter 7 (Continued) Command

Command Mode and Description

ip authentication mode eigrp asn md5

Enables MD5 authentication of EIGRP packets on an interface

ip authentication key-chain eigrp asn key_chain_name

Specifies the authentication key for EIGRP on an interface

distribute-list {access-list-number | name} {in | out} [interface-type interface-number]

Specifies an access list for filtering routing updates to/from the EIGRP topology table

distribute-list prefix prefix_list_name {in | out} [interface-type interfacenumber]

Specifies a prefix list for filtering routing updates to/from the EIGRP topology table

timers active-time [time-limit | disabled]

EIGRP config mode; sets the time limit for how long a route is in active state before becoming stuck-in-active

show ip route eigrp asn

User mode; displays all EIGRP routes in the IP routing table

show ip eigrp topology [as-number | [[ip-address] mask]] [active | all-links | pending | summary | zero-successors]

User mode; lists different parts of the EIGRP topology table, depending on the options used

show ip eigrp interfaces [interfacetype interface-number] [as-number]

User mode; lists EIGRP protocol timers and statistics per interface

show ip eigrp traffic [as-number]

User mode; displays EIGRP traffic statistics

show ip protocols

User mode; lists EIGRP timer settings, current protocol status, automatic summarization actions, and update sources

show ip eigrp asn neighbors

User mode; lists EIGRP neighbors

clear ip eigrp neighbors [ip-address | interface-type interface-number]

Enable mode; disables current neighbor relationships, removing topology table entries associated with each neighbor

clear ip route {network [mask] | *}

Enable mode; clears the routing table entries, which are then refilled based on the current topology table

show ip interface [type number]

User mode; lists many interface settings, including split horizon

eigrp log-neighbor-changes

EIGRP subcommand; displays log messages when neighbor status changes; enabled by default

246

Chapter 7: EIGRP

Table 7-7 summarizes the types of EIGRP packets and their purposes. Table 7-7

EIGRP Message Summary EIGRP Packet

Purpose

Hello

Identifies neighbors, exchanges parameters, and is sent periodically as a keepalive function

Update

Informs neighbors about routing information

Ack

Acknowledges Update, Query, and Response packets

Query

Asks neighboring routers to verify their route to a particular subnet

Reply

Sent by neighbors to reply to a Query

Goodbye

Used by a router to notify its neighbors when the router is gracefully shutting down

Memory Builders The CCIE Routing and Switching written exam, like all Cisco CCIE written exams, covers a fairly broad set of topics. This section provides some basic tools to help you exercise your memory about some of the broader topics covered in this chapter.

Fill In Key Tables from Memory Appendix G, “Key Tables for CCIE Study,” on the CD in the back of this book contains empty sets of some of the key summary tables in each chapter. Print Appendix G, refer to this chapter’s tables in it, and fill in the tables from memory. Refer to Appendix H, “Solutions for Key Tables for CCIE Study,” on the CD to check your answers.

Definitions Next, take a few moments to write down the definitions for the following terms: hello interval, full update, partial update, Route Tag field, Next Hop field, MD5, DUAL, Hold timer, K value, neighbor, adjacency, RTP, SRTT, RTO, Update, Ack, query, Reply, Hello, Goodbye, RD, FD, feasibility condition, successor route, feasible successor, input event, local computation, active, passive, going active, stuck-in-active, query scope, EIGRP stub router, limiting query scope, variance Refer to the glossary to check your answers.

Memory Builders

247

Further Reading Jeff Doyle’s Routing TCP/IP, Volume I, Second Edition, (Cisco Press) has several excellent examples of configuration, as well as several examples of the DUAL algorithm and the Active Query process. EIGRP Network Design Solutions, by Ivan Pepelnjak, contains wonderfully complete coverage of EIGRP. It also has great, detailed examples of the Query process.

Blueprint topics covered in this chapter: This chapter covers the following subtopics from the Cisco CCIE Routing and Switching written exam blueprint. Refer to the full blueprint in Table I-1 in the Introduction for more details on the topics covered in each chapter and their context within the blueprint. ■

Standard OSPF Area

■

Stub OSPF Area

■

Totally Stubby Area

■

Not-So-Stubby Area (NSSA)

■

Totally NSSA

■

LSA Types

■

Adjacency on Point-to-Point and Multiaccess Network Types

■

OSPF Graceful Restart

CHAPTER

8

OSPF This chapter covers OSPF, the only link-state routing protocol covered by the CCIE Routing and Switching exam blueprint. As with the other routing protocol chapters, this chapter includes most of the features, concepts, and commands related to OSPF. Chapter 9 “IGP Route Redistribution, Route Summarization, Default Routing, and Troubleshooting,” covers a few other details of OSPF, in particular, route redistribution, route filtering in redistribution, and route summarization.

“Do I Know This Already?” Quiz Table 8-1 outlines the major sections in this chapter and the corresponding “Do I Know This Already?” quiz questions. Table 8-1

“Do I Know This Already?” Foundation Topics Section-to-Question Mapping Foundation Topics Section

Questions Covered in This Section

OSPF Database Exchange

1–5

OSPF Design and LSAs

6–9

OSPF Configuration

10–12

Score

Total Score

In order to best use this pre-chapter assessment, remember to score yourself strictly. You can find the answers in Appendix A, “Answers to the ‘Do I Know This Already?’ Quizzes.” 1.

R1 has received an OSPF LSU from R2. Which of the following methods may be used by R1 to acknowledge receipt of the LSU from R2? a.

TCP on R1 acknowledges using the TCP Acknowledgement field.

b.

R1 sends back an identical copy of the LSU.

c.

R1 sends back an LSAck to R2.

d.

R1 sends back a DD packet with LSA headers whose sequence numbers match the sequence numbers in the LSU.

250

Chapter 8: OSPF

2.

3.

4.

Fredsco has an enterprise network with one core Frame Relay connected router, with a huband-spoke network of PVCs connecting to ten remote offices. The network uses OSPF exclusively. The core router (R-core) has all ten PVCs defined under multipoint subinterface s0/0.1. Each remote router also uses a multipoint subinterface. Fred, the engineer, configures an ip ospf network non-broadcast command under the subinterface on R-core and on the subinterfaces of the ten remote routers. Fred also assigns an IP address to each router from subnet 10.3.4.0/24, with R-core using the .100 address, and the remote offices using .1 through .10. Assuming all other related options are using defaults, which of the following would be true about this network? a.

The OSPF hello interval would be 30 seconds.

b.

The OSPF dead interval would be 40 seconds.

c.

The remote routers could learn all routes to other remote routers’ subnets, but only if R-core became the designated router.

d.

No designated router will be elected in subnet 10.3.4.0/24.

Which of the following interface subcommands, used on a multipoint Frame Relay subinterface, creates a requirement for a DR to be elected for the attached subnet? a.

ip ospf network point-to-multipoint

b.

ip ospf network point-to-multipoint non-broadcast

c.

ip ospf network non-broadcast

d.

None of these answers is correct.

The following routers share the same LAN segment and have the stated OSPF settings: R1: RID 1.1.1.1, hello 10, priority 3; R2: RID 2.2.2.2, hello 9, priority 4; R3, RID 3.3.3.3, priority 3; and R4: RID 4.4.4.4, hello 10, priority 2. The LAN switch fails, recovers, and all routers attempt to elect an OSPF DR and form neighbor relationships at the same time. No other OSPF-related parameters were specifically set. Which of the following is true about negotiations and elections on this LAN? a.

R1, R3, and R4 will expect Hellos from R2 every 9 seconds.

b.

R2 will become the DR but have no neighbors.

c.

R3 will become the BDR.

d.

R4’s dead interval will be 40 seconds.

e.

All routers will use R2’s hello interval of 9 once R2 becomes the designated router.

“Do I Know This Already?” Quiz

5.

6.

7.

251

Which of the following must be true in order for two OSPF routers that share the same LAN data link to be able to become OSPF neighbors? a.

Must be in the same area

b.

Must have the same LSRefresh setting

c.

Must have differing OSPF priorities

d.

Must have the same Hello timer, but can have different dead intervals

R1 is an OSPF ASBR that injects an E1 route for network 200.1.1.0/24 into the OSPF backbone area. R2 is an ABR connected to area 0 and to area 1. R2 also has an Ethernet interface in area 0, IP address 10.1.1.1/24, for which it is the designated router. R3 is a router internal to area 1. Enough links are up and working for the OSPF design to be working properly. Which of the following is true regarding this topology? (Assume no other routing protocols are running, and that area 1 is not a stub area.) a.

R1 creates a type 7 LSA and floods it throughout area 0.

b.

R3 will not have a specific route to 200.1.1.0/24.

c.

R2 forwards the LSA that R1 created for 200.1.1.0/24 into area 1.

d.

R2 will create a type 2 LSA for subnet 10.1.1.0/24 and flood it throughout area 0.

R1 is an OSPF ASBR that injects an E1 route for network 200.1.1.0/24 into the OSPF backbone area. R2 is an ABR connected to area 0 and to area 1. R2 also has an Ethernet interface in area 0, IP address 10.1.1.1/24, for which it is the designated router. R3 is a router internal to area 1. Enough links are up and working for the OSPF design to be working properly. Which of the following are true regarding this topology? (Assume no other routing protocols are running, and that area 1 is a totally NSSA area.) a.

R3 could inject internal routes into the OSPF domain.

b.

R3 will not have a specific route to 200.1.1.0/24.

c.

R2 forwards the LSA that R1 created for 200.1.1.0/24 into area 1.

d.

R2 will create a type 2 LSA for subnet 10.1.1.0/24 and flood it throughout area 0.

252

Chapter 8: OSPF

8.

9.

10.

The routers in area 55 all have the area 55 stub no-summary command configured under the router ospf command. OSPF has converged, with all routers in area 55 holding an identical link-state database for area 55. All IP addresses inside the area come from the range 10.55.0.0/16; no other links outside area 55 use addresses in this range. R11 is the only ABR for the area. Which of the following is true about this design? a.

The area is a stubby area.

b.

The area is a totally stubby area.

c.

The area is an NSSA.

d.

ABR R11 is not allowed to summarize the type 1 and 2 LSAs in area 55 into the 10.55.0.0/16 prefix due to the no-summary keyword.

e.

Routers internal to area 55 can have routes to specific subnets inside area 0.

f.

Routers internal to area 55 can have routes to E1, but not E2, OSPF routes.

R1 is an OSPF ASBR that injects an E1 route for network 200.1.1.0/24 into the OSPF backbone area. R2 is an ABR connected to area 0 and to area 1. R2 also has an Ethernet interface in area 0, IP address 10.1.1.1/24, for which it is the designated router. R3 is a router internal to area 1. Enough links are up and working for the OSPF design to be working properly. Which of the following are true regarding this topology? (Assume no other routing protocols are running, and that area 1 is not a stubby area.) a.

R3’s cost for the route to 200.1.1.0 will be the cost of the route as it was injected into the OSPF domain by R1, without considering any internal cost.

b.

R3’s cost for the route to 200.1.1.0 will include the addition of R3’s cost to reach R1, plus the external cost listed in the LSA.

c.

R3’s cost for the route to 10.1.1.0/24 will be the same as its cost to reach ABR R2.

d.

R3’s cost for the route to 10.1.1.0/24 will be the sum of its cost to reach ABR R2 plus the cost listed in the type 3 LSA created for 10.1.1.0/24 by ABR R2.

e.

It is impossible to characterize R3’s cost to 10.1.1.0/24 because R3 uses a summary type 3 LSA, which hides some of the costs.

R1 and R2 each connect via Fast Ethernet interfaces to the same LAN, which should be in area 0. R1’s IP address is 10.1.1.1/24, and R2’s is 10.1.1.2/24. The only OSPF-related configuration is as follows: hostname R1 router ospf 1 network 0.0.0.0 255.255.255.255 area 0 auto-cost reference-bandwidth 1000 ! hostname R2 router ospf 2 network 10.0.0.0 0.0.0.255 area 0

“Do I Know This Already?” Quiz

253

Which of the following statements are true about the configuration?

11.

12.

a.

The network command on R2 does not match IP address 10.1.1.2, so R2 will not attempt to send Hellos or discover neighbors on the LAN.

b.

The different process IDs in the router ospf command prevent the two routers from becoming neighbors on the LAN.

c.

R2 will become the DR as a result of having a cost of 1 associated with its Fast Ethernet interface.

d.

R1 and R2 could never become neighbors due to the difference in cost values.

e.

R1’s OSPF cost for its Fast Ethernet interface would be 10.

Which of the following are true about setting timers with OSPF? a.

The ip ospf dead-interval minimal hello-multiplier 4 interface subcommand sets the hello interval to 4 ms.

b.

The ip ospf dead-interval minimal hello-multiplier 4 interface subcommand sets the dead interval to 4 seconds.

c.

The ip ospf dead-interval minimal hello-multiplier 4 interface subcommand sets the hello interval to 250 ms.

d.

On all interfaces, the ip ospf hello-interval 30 interface subcommand changes the hello interval from 10 to 30.

e.

The ip ospf hello-multiplier 5 interface subcommand sets the dead interval to five times the then-current hello interval.

f.

Cisco IOS defaults the hello and dead intervals to 30/120 on interfaces using the OSPF nonbroadcast network type.

R1 has been configured for OSPF authentication on its fa0/0 interface as shown below. Which of the following is true about the configuration? interface fa0/0 ip ospf authentication-key hannah ip ospf authentication ip ospf message-digest-key 2 md5 jessie router ospf 2 area 0 authentication message-digest

a.

R1 will attempt simple-text authentication on the LAN with key hannah.

b.

R1 will attempt MD5 authentication on the LAN with key jessie.

c.

R2 will attempt OSPF type 2 authentication on fa0/0.

d.

R2 will attempt OSPF type 3 authentication on fa0/0.

254

Chapter 8: OSPF

Foundation Topics Link-state routing protocols define the content and structure of data that describes network topology, and define the processes by which routers exchange that detailed topology information. The name “link state” refers to the fact that the topology information includes information about each data link, along with each link’s current operational state. All the topological data together comprises the link-state database (LSDB). Each link-state router applies the Dijkstra algorithm to the database to calculate the current-best routes to each subnet. This chapter breaks down the OSPF coverage into three major sections. The first section details how the topology data is exchanged. The second section covers OSPF design and the contents of the LSDB, which comprises different types of link-state advertisements (LSAs). (The second section covers both design and the LSDB because the design choices directly impact which types of LSAs are forwarded into the differing parts of an OSPF network.) The third section covers the majority of the OSPF configuration details of OSPF for this chapter, although a few configuration topics are interspersed in the first two sections. NOTE This chapter addresses the functions of OSPF Version 2. It ignores OSPF Version 3 (RFC 2740), which was introduced primarily to support IPv6 and is covered in detail in Chapter 20, “IP Version 6.”

OSPF Database Exchange OSPF defines five different messages that routers can use to exchange LSAs. The process by which LSAs are exchanged does not change whether a single area or multiple areas are used, so this section will use a single OSPF area (area 0).

OSPF Router IDs Before an OSPF router can send any OSPF messages, it must choose a unique 32-bit dotteddecimal identifier called the OSPF router identifier (RID). Cisco routers use the following sequence to choose their OSPF RID, only moving on to the next step in this list if the previous step did not supply the OSPF RID: 1.

Use the router ID configured in the router-id id subcommand under router ospf.

2.

Use the highest numeric IP address on any currently “up and up” loopback interface.

3.

Use the highest numeric IP address on any currently “up and up” non-loopback interface.

OSPF Database Exchange

255

The sequence and logic are very simple, but some details are hidden in the sequence: ■

The interface from which the RID is taken does not have to be matched by an OSPF network command.

■

OSPF does not have to advertise a route to reach the RID’s subnet.

■

The RID does not have to be reachable per the IP routing table.

■

Steps 2 and 3 look at the then-current interface state to choose the RID when the OSPF process is started.

■

Routers consider changing the OSPF RID when the OSPF process is restarted, or when the RID is changed via configuration.

■

If a router’s RID changes, the rest of the routers in the same area will have to perform a new SPF calculation.

■

If the RID is configured with the router-id command, and the command remains unchanged, that router’s RID will never change.

For these reasons, many people set their RIDs with the router-id command and use an obvious numbering scheme to make it easy to identify a router by its RID.

Becoming Neighbors, Exchanging Databases, and Becoming Adjacent OSPF directly encapsulates the five different types of OSPF messages inside IP packets, using IP protocol 89, as listed in Table 8-2. Table 8-2

OSPF Messages Message

Description

Hello

Used to discover neighbors, bring a neighbor relationship to a 2-way state, and monitor a neighbor’s responsiveness in case it fails

Database Description (DD or DBD)

Used to exchange brief versions of each LSA, typically on initial topology exchange, so that a router knows a list of that neighbor’s LSAs

Link-State Request (LSR)

A packet that identifies one or more LSAs about which the sending router would like the neighbor to supply full details about the LSAs

Link-State Update (LSU)

A packet that contains fully detailed LSAs, typically sent in response to an LSR message

Link-State Acknowledgement (LSAck)

Sent to confirm receipt of an LSU message

256

Chapter 8: OSPF

These messages together allow routers to discover each other’s presence (Hello), learn which LSAs are missing from their LSDBs (DD), request and reliably exchange the LSAs (LSR/LSU), and monitor their neighbors for any changes in the topology (Hello). Note that the LSAs themselves are not OSPF messages—an LSA is a data structure, held inside a router’s LSDB, and exchanged inside LSU messages. When a particular data link first comes up, OSPF routers first become neighbors using the Hello message. At that point, they exchange topology information using the other four OSPF messages. Figure 8-1 outlines the overall process between two routers. Figure 8-1

Overview of OSPF LSDB Exchange Neighbor State

Neighbor State

Down

Down (R1 to R2 Link comes up ...)

RID 1.1.1.1

Init

RID 2.2.2.2

Init Hello, Seen [null], RID 1.1.1.1

R1 Hello, Seen [1.1.1.1], RID 2.2.2.2

2-way DR Election, if needed ExStart

Hello, Seen [1.1.1.1, 2.2.2.2], RID 1.1.1.1 Hello, DR=z.z.z.z

DR Election, if needed

DD (LSA Headers) DD (LSA Headers)

Exchange

R2

2-way

DD (LSA Headers)

ExStart Exchange

. . . Loading Full

LSR, LSU, LSAck (Full LSAs)

. . .

Loading Full

Figure 8-1 shows the overall message flow, along with the neighbor state on each router. An OSPF router keeps a state machine for each neighbor, listing the current neighbor state in the output of the show ip ospf neighbor command. These neighbor states change as the neighbors progress through their messaging; in this example, the neighbors settle into a full state, meaning fully adjacent, once the process is completed. The “Foundation Summary” section at the end of this chapter includes a reference table (Table 8-13) listing the neighbor states and their meanings. The next few sections explain the details behind the process shown in Figure 8-1.

OSPF Database Exchange

257

Becoming Neighbors: The Hello Process Hello messages perform three major functions: ■

Discover other OSPF-speaking routers on common subnets

■

Check for agreement on some configuration parameters

■

Monitor health of the neighbors to react if the neighbor fails

To discover neighbors, Cisco OSPF routers listen for multicast Hello messages sent to 224.0.0.5— the All OSPF Routers multicast address—on any interfaces that have been enabled for OSPF. The Hellos are sourced from that router’s primary IP address on the interface—in other words, Hellos are not sourced from secondary IP addresses. (OSPF routers will advertise secondary IP addresses, but they will not send Hellos from those IP addresses, and never form neighbor relationships using secondary addresses.) Furthermore, OSPF neighbors will become fully adjacent if one or both of the neighbors are using unnumbered interfaces for the connection between them. After two routers discover each other by receiving Hellos from the other router, the routers perform the following parameter checks based on the receive Hellos: ■

Must pass the authentication process

■

Must be in the same primary subnet, including same subnet mask

■

Must be in the same OSPF area

■

Must be of the same area type (stub, NSSA, and so on)

■

Must not have duplicate RIDs

■

OSPF Hello and Dead timers must be equal

If any of these items do not match, the two routers simply do not form a neighbor relationship. Also of note is one important item that does not have to match: the OSPF process ID (PID), as configured in the router ospf process-id command. Also, the MTU must be equal for the DD packets to be successfully sent between neighbors, but this parameter check is technically not part of the Hello process. The third important function for a Hello is to maintain a heartbeat function between neighbors. The neighbors send Hellos every hello interval; failure to receive a Hello within the longer dead interval causes a router to believe that its neighbor has failed. The hello interval defaults to 10 seconds on LAN interfaces and 30 seconds on T1 and slower WAN interfaces; the dead interval defaults to four times the hello interval. Example 8-1 lists some basic OSPF command output related to the neighbor establishment with Hellos, and the hello and dead intervals.

258

Chapter 8: OSPF

Example 8-1

Hello Mismatches and Basic Neighbor Parameters

! Below, debug messages show that this router disagrees with the hello and dead ! intervals on router 10.1.111.4; The “C” and “R” mean “configured” and “received,” ! respectively, meaning that this router uses 30/120 for hello/dead, and the other ! router is trying to use 10/40. R1# debug ip ospf hello OSPF hello events debugging is on Jan 12 06:41:20.940: OSPF: Mismatched hello parameters from 10.1.111.4 Jan 12 06:41:20.940: OSPF: Dead R 40 C 120, Hello R 10 C 30 255.255.255.0

Mask R 255.255.255.0 C

! Below, R1’s hello and dead intervals are listed for the same interface. R1# show ip ospf int s 0/0.100 Serial0/0.100 is up, line protocol is up Internet Address 10.1.111.1/24, Area 0 Process ID 1, Router ID 1.1.1.1, Network Type NON_BROADCAST, Cost: 64 Transmit Delay is 1 sec, State DR, Priority 1 Designated Router (ID) 1.1.1.1, Interface address 10.1.111.1 No backup designated router on this network Timer intervals configured, Hello 30, Dead 120, Wait 120, Retransmit 5 ! Lines omitted for brevity ! Below, R1 shows a neighbor on S0/0.100, in the full state, meaning the routers ! have completed LSDB exchange. Note the current Dead timer counts down, in this ! case from 2 minutes; the value of 1:58 means R1 last received a Hello from ! neighbor 10.1.111.6 two seconds ago. R1# sh ip ospf neighbor 6.6.6.6 Neighbor 6.6.6.6, interface address 10.1.111.6 In the area 0 via interface Serial0/0.100 Neighbor priority is 0, State is FULL, 8 state changes DR is 10.1.111.1 BDR is 0.0.0.0 Poll interval 120 Options is 0x42 Dead timer due in 00:01:58 Neighbor is up for 00:17:22 ! Lines omitted for brevity

Flooding LSA Headers to Neighbors When two routers hear Hellos, and the parameter check passes, they do not immediately send packets holding the LSAs. Instead, each router creates and sends Database Description (DD, or sometimes called DBD) packets, which contain the headers of each LSA. The headers include enough information to uniquely identify each LSA. Essentially, the routers exchange a list of all the LSAs they each know about; the next step in the process is letting a router request a new copy of any old or unknown LSAs. The DD messages use an OSPF-defined simple error-recovery process. Each DD packet, which may contain several LSA headers, has an assigned sequence number. The receiver acknowledges a received DD packet by sending an identical DD packet back to the sender. The sender uses a window size of one packet, then waits for the acknowledgement before sending the next DD packet.

OSPF Database Exchange

259

Database Descriptor Exchange: Master/Slave Relationship As a neighbor relationship forms between two routers (specifically, at the ExStart stage of the neighborship), the neighbors determine which router is to be the master and which is to be the slave during the database exchange between them. The router with the higher RID becomes the master and initiates the database exchange. At that point, the master begins sending DD packets to the slave, and the slave acknowledges them as they are received. Only the master can increment sequence numbers in the DD exchange process. Requesting, Getting, and Acknowledging LSAs Once all LSA headers have been exchanged using DD packets, each neighboring router has a list of LSAs known by the neighbor. Using that knowledge, a router needs to request a full copy of each LSA that is missing from its LSDB. To know whether a neighbor has a more recent copy of a particular LSA, a router looks at the sequence number of the LSA in its LSDB and compares it to the sequence number of that same LSA learned from the DD packet. Each LSA’s sequence number is incremented every time the LSA changes. So, if a router received (via a DD packet) an LSA header with a later sequence number for a particular LSA (as compared with the LSA in the LSDB), that router knows that the neighbor has a more recent LSA. For example, R1 sent R2 an LSA header for the type 1 LSA that describes R1 itself, with sequence number 0x80000004. If R2’s database already held that LSA, but with a sequence number of 0x80000003, then R2 would know that it needs to ask R1 to send the latest copy (sequence number 0x80000004) of that LSA. NOTE New LSAs begin with sequence number 0x80000001, increase, and then wrap back to 0x7FFFFFFF. If the LSA made it to sequence number 0x80000000, the LSA must be reflooded throughout the network.

Routers use Link-State Request (LSR) packets to request one or more LSAs from a neighbor. The neighboring router replies with Link-State Update (LSU) packets, which hold one or more full LSAs. As shown in Figure 8-1, both routers sit in a loading state while the LSR/LSA process continues. Once the process is complete, they settle into a full state, which means that the two routers should have fully exchanged their databases, resulting in identical copies of the LSDB entries for that area on both routers. The LSR/LSA process uses a reliable protocol that has two options for acknowledging packets. First, an LSU can be acknowledged by the receiver of the LSU simply repeating the exact same LSU back to the sender. Alternatively, a router can send back an LSAck packet to acknowledge the packet, which contains a list of acknowledged LSA headers. At the end of the process outlined in Figure 8-1, two neighbors have exchanged their LSDBs. As a result, their LSDBs should be identical. At this point, they can each independently run the Dijkstra Shortest Path First (SPF) algorithm to calculate the best routes from their own perspectives.

260

Chapter 8: OSPF

Designated Routers on LANs OSPF optimizes the LSA flooding process on multiaccess data links by using the concept of a designated router (DR). Without the concept of a DR, each pair of routers that share a data link would become fully adjacent neighbors. Each pair of routers would directly exchange their LSDBs with each other as shown in Figure 8-1. On a LAN with only six routers, without a DR, 15 different pairs of routers would exist, and 15 different instances of full database flooding would occur. OSPF uses a DR (and backup DR, or BDR) on a LAN or other multiaccess network. The flooding occurs through the DR, significantly reducing the unnecessary exchange of redundant LSAs. NOTE DRs have one other major function besides improving the efficiency of LSA flooding process. They also create a type 2 LSA that represents the subnet. LSA types are covered in the next major section, “OSPF Design and LSAs.” The next section goes through the basics of the DR/BDR process on LANs, which is followed by coverage of options of OSPF network types and how they impact OSPF flooding on Frame Relay links. Designated Router Optimization on LANs Figure 8-2 depicts the DR flooding optimization that occurs with sending DD packets over a LAN. Figure 8-2

DR Optimization on a LAN Steps: 1. R1 sends DD to all DR Mcast address (224.0.0.6) 2. DR sends unicast acknowledgement by repeating same DD 3. DR sends DD with same info as multicast to all SPF routers (224.0.0.5)

DROther (1.1.1.1) 1 R1

Legend: DR BDR DROther

BDR (7.7.7.7) DD, to 224.0.0.6 (All_DR)

S1

Designated Router Backup DR Neither the DR or BDR

3

DD, to 224.0.0.5 (All_SPF)

2

DD, in Acknowledgement, to R1, Unicast

R2 DROther (2.2.2.2)

S2 DR (8.8.8.8)

OSPF Database Exchange

261

Routers that are not the DR (including the BDR) send DDs to the DR by sending them to multicast address 224.0.0.6, the All OSPF DR Routers multicast address. The DR then acknowledges the DDs with a unicast DD (Step 2 in Figure 8-2). The DR then floods a new DD packet to all OSPF routers (multicast address 224.0.0.5). Figure 8-2 shows the three main steps, but the non-DR routers also need to acknowledge the DD packet sent in Step 3. Typically, the acknowledgment occurs by the other routers each replying with a unicast DD packet. NOTE In topologies without a DR, the DD and LSU packets are typically sent to the 224.0.0.5 All OSPF Routers multicast IP address.

Example 8-2 shows the output of a show ip ospf neighbor command on R1 from Figure 8-2. Note that R1 is in a full state with S2, which is the DR, with OSPF RID 8.8.8.8. R1 is also in a full state with S1, the BDR, OSPF RID 7.7.7.7. However, R1 is in a 2WAY state with R2, RID 2.2.2.2. Example 8-2

The show ip ospf neighbor Command

R1# sh ip ospf neighbor fa 0/0 Neighbor ID

State

Dead Time

Address

Interface

2.2.2.2

Pri 1

2WAY/DROTHER

00:00:35

10.1.1.2

FastEthernet0/0

7.7.7.7

1

FULL/BDR

00:00:38

10.1.1.3

FastEthernet0/0

8.8.8.8

1

FULL/DR

00:00:34

10.1.1.4

FastEthernet0/0

When a DR is used on a link, routers end up as DR, BDR, or neither; a router that is neither DR or BDR is called a DROther router. The DR and BDR form full adjacencies with all other neighbors on the link, so they reach a full state once the database exchange process is complete. However, two neighbors that are both DROthers do not become fully adjacent—they stop at the 2WAY state, as shown in Example 8-2. Stopping at the 2WAY state between two DROther routers is normal; it simply means that the Hello parameter-match check worked, but the neighbors do not need to proceed to the point of exchanging DD packets, because they do not need to when a DR is present. To describe the fact that some neighbors do not directly exchange DD and LSU packets, OSPF makes a distinction between the terms neighbors and adjacent, as follows: ■

Neighbors—Two routers that share a common data link, that exchange Hello messages, and the Hellos must match for certain parameters.

■

Adjacent (fully adjacent)—Two neighbors that have completed the process of fully exchanging DD and LSU packets directly between each other.

262

Chapter 8: OSPF

Note that although DROther routers do not exchange DD and LSU packets directly with each other, like R1 and R2 in Figure 8-2, the DROther routers do end up with an identical copy of the LSDB entries by exchanging them with the DR.

DR Election on LANs As noted in Figure 8-1, if a DR is elected, the election occurs after the routers have become neighbors, but before they send DD packets and reach the ExStart neighbor state. When an OSPF router reaches the 2-way state with the first neighbor on an interface, it has already received at least one Hello from that neighbor. If the Hello messages state a DR of 0.0.0.0—meaning none has been elected—the router waits before attempting to elect a DR. This typically occurs after a failure on the LAN. OSPF routers wait with the goal of giving all the routers on that subnet a chance to finish initializing after a failure so that all the routers can participate in the DR election—otherwise, the first router to become active would always become the DR. (The time period is called the OSPF wait time, which is set to the same value as the Dead timer.) However, if the received Hellos already list the DR’s RID, the router does not have to wait before beginning the election process. This typically occurs when one router lost its connection to the LAN, but other routers remained and continued to work. In this case, the newly-connected router does not attempt to elect a new DR, assuming the DR listed in the received Hello is indeed the current DR. The election process allows for the possibility of many different scenarios for which routers may and may not become the DR or BDR. Generally speaking, the following rules govern the DR/BDR election process: ■

Any router with its OSPF priority set to between 1–255 inclusive can try to become DR by putting its own RID into the DR field of its sent Hellos.

■

Routers examine received Hellos, looking at other routers’ priority settings, RIDs, and whether each neighbor claims to want to become the DR.

■

If a received Hello implies a “better” potential DR, the router stops claiming to want to be DR and asserts that the better candidate should be the DR.

■

The first criteria for “better” is the router with the highest priority.

■

If the priorities tie, the router with the higher RID is better.

■

The router not claiming to be the DR, but with the higher priority (or higher RID, in case priority is a tie) becomes the BDR.

OSPF Database Exchange

263

■

If a new router arrives after the election, or an existing router improves its priority, it cannot preempt the existing DR and take over as DR (or as BDR).

■

When a DR is elected, and the DR fails, the BDR becomes DR, and a new election is held for a new BDR.

After the DR is elected, LSA flooding continues as illustrated previously in Figure 8-2.

Designated Routers on WANs and OSPF Network Types Using a DR makes good sense on a LAN because it improves LSA flooding efficiency. Likewise, not using a DR on a point-to-point WAN link also makes sense, because with only two routers on the subnet, there is no inefficiency upon which to improve. However, on nonbroadcast multiaccess (NBMA) networks, arguments can be made regarding whether a DR is helpful. So, OSPF includes several options that include a choice of whether to use a DR on WAN interfaces. Cisco router interfaces can be configured to use, or not use, a DR, plus a couple of other key behaviors, based on the OSPF network type for each interface. The OSPF network type determines that router’s behavior regarding the following: ■

Whether the router tries to elect a DR on that interface

■

Whether the router must statically configure a neighbor (with the neighbor command), or find neighbors using the typical multicast Hello packets

■

Whether more than two neighbors should be allowed on the same subnet

For instance, LAN interfaces default to use an OSPF network type of broadcast. OSPF broadcast networks elect a DR, use Hellos to dynamically find neighbors, and allow more than two routers to be in the same subnet on that LAN. For HDLC and PPP links, OSPF uses a network type of point-to-point, meaning that no DR is elected, only two IP addresses are in the subnet, and neighbors can be found through Hellos. Table 8-3 summarizes the OSPF interface types and their meanings. Note that the interface type values can be set with the ip ospf network type interface subcommand; the first column in the table lists the exact keyword according to this command. Also, for cases in which a DR is not elected, all routers that become neighbors also attempt to become adjacent by the direct exchange of DD, LSR, and LSU packets.

264

Chapter 8: OSPF

OSPF Network Types

Table 8-3

Interface Type

Uses DR/ BDR?

Default Hello Interval

Requires a neighbor Command?

More than Two Hosts Allowed in the Subnet?

Broadcast

Yes

10

No

Yes

Point-to-point1

No

10

No

No

Nonbroadcast2 (NBMA)

Yes

30

Yes

Yes

Point-to-multipoint

No

30

No

Yes

Point-to-multipoint nonbroadcast

No

30

Yes

Yes

Loopback

No

—

—

No

1

Default on Frame Relay point-to-point subinterfaces.

2

Default on Frame Relay physical and multipoint subinterfaces.

Caveats Regarding OSPF Network Types over NBMA Networks When configuring OSPF over Frame Relay, the OSPF network type concept can become a bit troublesome. In fact, many CCIE Routing and Switching lab preparation texts and lab books focus on the variety of combinations of OSPF network types used with Frame Relay for various interfaces/subinterfaces. The following list contains many of the key items you should check when looking at an OSPF configuration over Frame Relay, when the OSPF network types used on the various routers do not match: ■

Make sure the default Hello/Dead timers do not cause the Hello parameter check to fail. (See Table 8-3 for the defaults for each OSPF network type.)

■

If one router expects a DR to be elected, and the other does not, the neighbors may come up, and full LSAs be communicated. However, show command output may show odd information, and next-hop routers may not be reachable. So, make sure all routers in the same NBMA subnet use an OSPF network type that either does use a DR or does not.

■

If a DR is used, the DR and BDR must have a permanent virtual circuit (PVC) to every other router in the subnet. If not, not all routers will be able to learn routes, because the DR must forward the DD and LSU packets to each of the other routers. Routers that do not have a PVC to every other router should not be permitted to become a DR/BDR.

■

If one router requires a static neighbor command, typically the other router on the other end of the PVC does not require a neighbor command. For clarity, however, it is better to configure neighbor commands on both routers.

OSPF Database Exchange

265

Two very simple options exist for making OSPF work over Frame Relay—both of which do not require a DR and do not require neighbor commands. If the design allows for the use of pointto-point subinterfaces, use those, take the default OSPF network type of point-to-point, and no additional work is required. If multipoint subinterfaces are needed, or if the configuration must not use subinterfaces, adding the ip ospf network point-to-multipoint command on all the routers works, without requiring additional effort to manually define neighbors or worry about which router becomes the DR. Example of OSPF Network Types and NBMA On NBMA networks with an OSPF network type that requires that a DR be elected, you must take care to make sure the correct DR is elected. The reason is that the DR and BDR must each have a PVC connecting it to all the DROther routers—otherwise, LSA flooding will not be possible. So, with partial meshes, the election should be influenced by configuring the routers’ priority and RIDs such that the hub site of a hub-and-spoke partial mesh becomes the DR. Figure 8-3 shows an example network for which R1 should be the only router allowed to become DR or BDR. Figure 8-3

Network Used in the Frame Relay Priority and Network Type Example 10.3.1.3/23 E0/0

R3

S0/0.1 Mpt

10.1.111.0/24 FR

10.3.1.4/23 E0/0

R4

S0/0.1 P-P

RID 1.1.1.1 R1

10.5.1.5/24 E0/0

R5

S0.1 Mpt

S0/0 Phy.

S0.1 Mpt 10.6.1.6/24 E0/0

R6

Example 8-3 depicts the following scenarios relating to DR election in Figure 8-3: ■

The R1, R3, and R5 configuration is correct for operating with default OSPF network type nonbroadcast in a partial mesh.

■

R6 has omitted the ip ospf priority interface subcommand, causing it to inadvisably become the DR.

■

R4 will be used as an example of what not to do, in part to point out some interesting facts about OSPF show commands.

266

Chapter 8: OSPF

NOTE Figure 8-3 and Example 8-3 do not depict a suggested design for Frame Relay and OSPF. With this topology, using point-to-point subinterfaces in all cases, using four small (/30) subnets, and defaulting to OSPF network type point-to-point would work well. Such a design, however, would not require any thought regarding the OSPF network type. So, this example is purposefully designed to provide a backdrop from which to show how the OSPF network types work. Example 8-3 shows only the nondefault OSPF configuration settings; also, the routers have an obvious RID numbering scheme (1.1.1.1 for R1, 2.2.2.2 for R2, and so on). Example 8-3

Setting Priority on NBMA Networks

! R1 configuration—the neighbor commands default to a priority value of 0, ! meaning R1’s perception of that neighbor is priority 0. router ospf 1 l o g - a d j a c e nc y - c h a n g e s d e t a i l n etwork 0.0.0.0 2 55.255.255.255 a rea 0 neigh bor 10.1.111.3 neigh bor 10.1.111.4 neighbor 10.1.111.5 nei ghbor 10.1.111.6 ! R3 configuration—R3’s interface priority is set to 0; R1 will use the higher ! of R3’s announced priority 0 (based on R3’s ip ospf priority interface ! subcommand) and the priority value on R1’s neighbor command, which defaulted ! to 0. So, R3 will not ever become a DR/BDR. interface Serial0/0.1 multipoint i p a d d r e s s 1 0 . 1 . 11 1 . 3 2 5 5 . 2 5 5 . 2 5 5 . 0 i p o s pf p r i o r i t y 0 frame-relay interface-dlci 100 ! R4 configuration—note from Figure 8-3 that R4 is using a point-to-point ! subinterface, with all defaults. This is not a typical use of a point-to-point ! subinterface, and is shown to make a few points later in the example. router osp f 1 n e t w o r k 0 . 0. 0 . 0 2 5 5 . 2 5 5 . 2 5 5 . 2 5 5 a r e a 0 ! R5’s configuration is equivalent to R3 in relation to the OSPF network type ! and its implications. interface Serial0.1 multipoint i p a d d r e s s 1 0 . 1 . 1 1 1. 5 2 5 5 . 2 5 5 . 2 5 5 . 0 ip o spf priority 0 frame-relay interface-dlci 100 ! router ospf 1 n e t w o r k 0 . 0 . 0 . 0 2 5 5 .2 5 5 . 2 5 5 . 2 5 5 a r e a 0 ! R6 configuration—R6 forgot to set the interface priority with the ip ospf ! priority 0 command, defaulting to priority 1. r outer ospf 1 network 0.0.0.0 255.255.255.255 area 0

OSPF Database Exchange

Setting Priority on NBMA Networks (Continued)

Example 8-3

! Below, the results of R6’s default interface priority of 1—R6, with RID ! 6.6.6.6, and an announced priority of 1, wins the DR election. Note that the ! command is issued on R1. R1# show ip ospf neighbor Neighbor ID

State

Dead Time

Address

Interface

6.6.6.6

Pri 1

FULL/DR

00:01:52

10.1.111.6

Serial0/0

3.3.3.3

0

FULL/DROTHER

00:01:46

10.1.111.3

Serial0/0

N/A

0

ATTEMPT/DROTHER

5.5.5.5

0

FULL/DROTHER

—

10.1.111.4

00:01:47

10.1.111.5

Serial0/0 Serial0/0

! Next, R1’s neighbor command was automatically changed to “priority 1” based on ! the Hello, with priority 1, that R1 received from R6. To prevent this dynamic ! reconfiguration, you could add an ip ospf priority 0 command under R6’s s0/0.1 ! interface. R1# show run | beg router ospf 1 router ospf 1 n e t w o r k 0 . 0 . 0 . 0 2 5 5 . 2 5 5 .2 5 5 . 2 5 5 a r e a 0 neighbor 10.1.111.6 priority 1 n e i gh b o r 1 0 . 1 . 1 1 1 . 3 n e i g h b o r 10 . 1 . 1 1 1 . 4 n e i g h b o r 1 0 . 1 . 11 1 . 5 ! lines omitted for brevity ! Below, R4 is OSPF network type “point to point,” with Hello/dead of 10/40. ! R1’s settings, based on Table 8-3, would be nonbroadcast, 30/120. R4# show ip ospf int s 0/0.1 Serial0/0.1 is up, line protocol is up Internet Address 10.1.111.4/24, Area 0 Process ID 1, Router ID 4.4.4.4, Network Type POINT_TO_POINT, Cost: 1562 Transmit Delay is 1 sec, State POINT_TO_POINT, Timer intervals configured, Hello 10, Dead 40, Wait 40, Retransmit 5 ! lines omitted for brevity ! Below, R4 changes its network type to yet a different value, one that expects ! neighbor commands, but does not expect a DR to be used. R4# conf t Enter configuration commands, one per line.

End with CNTL/Z.

R4(config)# int s 0/0.1 R4(config-subif)# ip ospf network point-to-multipoint non-broadcast ! Next, R1 and R4 become neighbors now that the Hello parameters match. Note that ! R1 believes that R4 is DROther. R1# show ip ospf neighbor Neighbor ID

Pri

State

Dead Time

Address

Interface

00:01:56

10.1.111.4

Serial0/0

! lines omitted for brevity 4.4.4.4

1

FULL/DROTHER

! Below, R4 agrees it is in a full state with R1, but does not list R1 as DR, ! because R4 is not using the concept of a DR at all due to R4’s network type. R4# sh ip ospf neigh Neighbor ID 1.1.1.1

Pri 0

State FULL/ —

Dead Time 00:01:42

Address 10.1.111.1

Interface Serial0/0.1

267

268

Chapter 8: OSPF

The first and most important point from Example 8-3 is the actual behavior of the two ways to set the priority in the example. The Cisco IOS Configuration Guide at Cisco.com states that the OSPF neighbor command defines the priority of the neighbor. However, in practice, a router’s neighbor priority setting is compared with the priority inside the Hello it receives from that neighbor— and the larger of the two values is used. In this example, R1’s neighbor 10.1.111.6 command (with default priority of 0) was overridden by R6’s Hello, which was based on R6’s default OSPF interface priority of 1. So, during DR election, R1 and R6 tied on OSPF priority, and R6 won due to its larger (6.6.6.6 versus 1.1.1.1) RID. R1 even automatically changed its neighbor command dynamically to neighbor 10.1.111.6 priority 1 to reflect the correct priority for R6. Also note that, although neighbors must be statically configured for some network types, the neighbor command needs to be configured on only one router. R3 and R5, with correct working configurations, did not actually need a neighbor command. Finally, it might seem that all is now fine between R1 and R4 by the end of the example, but even though the neighbors are fully adjacent, R4 cannot route packets to R3, R5, or R6 over the Frame Relay network. For instance, R5 could have some routes that point to 10.1.111.4 (R4’s Frame Relay IP address) as the next hop. However, because R5 is using a multipoint subinterface, R5 will not know what PVC to use to reach 10.1.111.4. (Chapter 6, “IP Forwarding (Routing),” covers how Frame Relay mapping occurs, and the logic used on multipoint and point-to-point subinterfaces.) In this case, the routers with multipoint subinterfaces would need to add frame-relay map commands; for example, R5 would need a frame-relay map ip 10.1.111.4 100 broadcast command, causing packets to next-hop 10.1.111.4 to go over DLCI 100 to R1, which would then route the packet on to R4. Keep in mind that R4’s configuration is not a recommended configuration.

SPF Calculation So far, this chapter has covered a lot of ground related to the exchange of LSAs. Regardless of the OSPF network type and whether DRs are used, once a router has new or different information in its LSDB, it uses the Dijkstra SPF algorithm to examine the LSAs in the LSDB and derive the mathequivalent of a figure of a network. This mathematical model has routers, links, costs for each link, and the current (up/down) status of each link. Figure 8-4 represents the SPF model of a sample network. Figure 8-4

Single-Area SPF Calculation: Conceptual View Route S2 – R1 – R5 : Cost 1 + 100 + 10 = 111 Route S2 – R2 – R5 : Cost 1 + 64 + 10 = 75

R1

C 100

C1

10.5.15.0/24

10.1.1.0/24

C 100 10.5.1.0/24 C 10

C1

C1

VLAN 1

10.5.25.0/24

R5

C 50

C 64

R2

S2

OSPF Design and LSAs

269

Humans can easily see the conclusion that the SPF algorithm will reach, even though the algorithm itself is fairly complicated. SPF on a router finds all possible routes to each subnet, adds the cost for each outgoing interface in that route, and then picks the path with the least cost. OSPF then places those least (shortest) cost routes into the routing table. For example, S2 calculates two possible routes to subnet 10.5.1.0/24, with the better route being out S2’s VLAN 1 interface, with R2 as the next-hop router. Also note in Figure 8-4 that the cost values are per interface, and it is each outgoing interface’s cost that SPF adds to come up with the total cost of the route.

Steady-State Operation Even after a network has stabilized, all routers in the same area have the exact same LSAs, and each router has chosen its best routes using SPF, the following is still true of routers running OSPF: ■

Each router sends Hellos, based on per-interface hello intervals.

■

Each router expects to receive Hellos from neighbors within the dead interval on each interface; if not, the neighbor is considered to have failed.

■

Each router originally advertising an LSA refloods each LSA (after incrementing its sequence number by 1) based on a per-LSA Link-State Refresh (LSRefresh) interval (default 30 minutes).

■

Each router expects to have its LSA refreshed within each LSA’s Maxage timer (default 60 minutes).

OSPF Design and LSAs This section covers two major topics: ■

OSPF design

■

OSPF LSA types

Although these might seem to be separate concepts, most OSPF design choices directly impact the LSA types in a network and impose restrictions on which neighbors may exchange those LSAs. This section starts with an OSPF design and terminology review, and then moves on to LSA types. Toward the end of the section, OSPF area types are covered, including how each variation changes how LSAs flow through the different types of OSPF stubby areas.

270

Chapter 8: OSPF

OSPF Design Terms OSPF design calls for grouping links into contiguous areas. Routers that connect to links in different areas are Area Border Routers (ABRs). ABRs must connect to area 0, the backbone area, and one or more other areas as well. Autonomous System Boundary Routers (ASBRs) inject routes external to OSPF into the OSPF domain, having learned those routes from wide-ranging sources from the Border Gateway Protocol (BGP) on down to simple redistribution of static routes. Figure 8-5 shows the terms in the context of a simple OSPF design. OSPF Design Terminology

Figure 8-5

Area x Stub Network Internal Router

Area Border Router

Backbone Router

Autonomous System Border Router

ABR1

BB2

ASBR1

R1 Transit Network

Another Routing Protocol

R2

R3

Area 0 Backbone Area

Area y

BB2 Backbone Router

R4 ABR2 R5

Area Border Router

Networks can use a single OSPF area, but using OSPF areas helps speed convergence and reduce overhead in an OSPF network. Using areas provides the following benefits: ■

Generally smaller per-area LSDBs, requiring less memory.

■

Faster SPF computation due to the sparser LSDB.

■

A link failure in one area only requires a partial SPF computation in other areas.

■

Routes may only be summarized at ABRs (and ASBRs); having areas allows summarization, again shrinking the LSDB and improving SPF calculation performance.

When comparing the use of one area versus using many areas, the number of routers or subnets does not shrink, but the size of the LSDB on most routers should shrink. The LSDB shrinks because an ABR does not pass denser and more detailed type 1 and 2 LSAs from one area to

OSPF Design and LSAs

271

another—instead, it passes type 3 summary LSAs. LSA types 1 and 2 can be thought of as the detailed topology information that causes most of the computing-intensive parts of the SPF algorithm; by representing these detailed type 1 and 2 LSAs in a different way in other areas, OSPF achieves its goal of reducing the effects of SPF. OSPF Path Selection Process OSPF has specific rules for selecting a path that crosses areas. Before studying the details of OSPF LSAs, it might help at this point to understand those rules: ■

Take the shortest path to area 0.

■

Take the shortest path across area 0 without traversing a nonzero area.

■

Take the shortest path to the destination without traversing area 0.

Note that these conditions can result in both asymmetric routing and suboptimal routing across multiarea OSPF networks. For example, if the shortest path to a destination in area 0 is not also the least-cost path, OSPF behaves more like distance vector protocols than the link-state protocol that it is, which can cause headaches in both design and troubleshooting.

LSA Types and Network Types Table 8-4 lists the LSA types and their descriptions for reference; following the table, each type is explained in more detail, in the context of a working network. OSPF LSA Types

Table 8-4

LSA Type

Common Name

1

Router

One per router, listing RID and all interface IP addresses. Represents stub networks as well.

2

Network

One per transit network. Created by the DR on the subnet, and represents the subnet and the router interfaces connected to the subnet.

3

Net Summary

Created by ABRs to represent one area’s type 1 and 2 LSAs when being advertised into another area. Defines the links (subnets) in the origin area, and cost, but no topology data.

4

ASBR Summary

Like a type 3 LSA, except it advertises a host route used to reach an ASBR.

5

AS External

Created by ASBRs for external routes injected into OSPF.

6

Group Membership

Defined for MOSPF; not supported by Cisco IOS.

7

NSSA External

Created by ASBRs inside an NSSA area, instead of a type 5 LSA.

8

External Attributes

Not implemented in Cisco routers.

9–11

Opaque

Used as generic LSAs to allow for easy future extension of OSPF; for example, type 10 has been adapted for MPLS traffic engineering.

Description

272

Chapter 8: OSPF

Before diving into the coverage of LSA types, two more definitions are needed: ■

Transit network—A network over which two or more OSPF routers have become neighbors and elected a DR, so traffic can transit from one to the other.

■

Stub network—A subnet on which a router has not formed any neighbor relationships.

Now on to the LSA types! LSA Types 1 and 2 Each router creates and floods a type 1 LSA for itself. These LSAs describe the router, its interfaces (in that area), and a list of neighboring routers (in that area) on each interface. The LSA itself is identified by a link-state ID (LSID) equal to that router’s RID. Type 2 LSAs represent a transit subnet for which a DR has been elected. The LSID is the DR’s interface IP address on that subnet. Note that type 2 LSAs are not created for subnets on which no DR has been elected. Armed with an LSDB with all the type 1 and 2 LSAs inside an area, a router’s SPF algorithm should be able to create a topological graph of the network, calculate the possible routes, and finally choose the best routes. For example, Figure 8-6 shows a sample internetwork that is used in several upcoming examples. Figure 8-7 shows a graphical view of the type 1 and type 2 LSAs created in area 3. Figure 8-6

Network Used in LSA Examples Stub 10.3.2.0/23

Area 3

RID 10.3.3.33 C 1

R33 C1

RID 3.3.3.3

Transit, with DR 10.3.0.0/23 E0/0 C 10

C 64 Transit, no DR 10.3.13.0/24

R3

RID 1.1.1.1

C 64

Area 4

RID 4.4.4.4 10.4.14.0/24

10.4.1.4/23 E0/0

C 1562

R4

S1 ASBR RID 7.7.7.7

R1

Area 5

10.5.15.0/24

C 100

10.1.1.0/24 RID 8.8.8.8

C 100 10.5.25.0/24

10.5.1.5/24 E0/0

R5 RID 5.5.5.5

C 50

C 64 RID 2.2.2.2

S2

R2

Area 0

Externals: 192.168.1.0/24 (E1) 192.168.2.0/24 (E2)

OSPF Design and LSAs

Figure 8-7

273

Graph of Type 1 and 2 LSAs for Area 3 E0/0 Stub 10.3.2.33/23 Cost 1

LSA Type 1 RID 10.3.3.33

Represents R33

E0/0 10.3.1.33 To Transit with DR 10.3.1.3 Cost 1

To RID 10.3.3.33 Interface 10.3.1.33

LSA Type 2 DR 10.3.1.3 Subnet 10.3.0.0/23

To RID 3.3.3.3 Interface 10.3.1.3 E0/0 10.3.1.3 Transit with DR 10.3.1.3 Cost 10

Represents Transit Net 10.3.0.0/23

LSA Type 1 RID 3.3.3.3

S0/0.1 Stub 10.3.13.2/24 To RID 1.1.1.1 Cost 64 S0/0.3 Stub 10.3.13.1/24 To RID 3.3.3.3 Cost 64

Represents R3

LSA Type 1 RID 1.1.1.1

Represents R1

For subnets without a DR, the type 1 LSAs hold enough information for the SPF algorithm to create the math model of the topology. For example, R1 and R3 use point-to-point subinterfaces, and the OSPF point-to-point network type. SPF can match up the information shown in the type 1 LSAs for R1 and R3 in Figure 8-7 to know that the two routers are connected. For transit networks with DRs, OSPF uses a type 2 LSA to model the subnet as a node in the SPF mathematical model. Because the SPF process treats the type 2 LSA as a node in the graph, this LSA is sometimes called a pseudonode. The type 2 LSA includes references to the RIDs of all routers that are currently neighbors of the DR on that subnet. That information, combined with the type 1 LSAs for each router connected to the subnet represented by the type 2 LSA, allows SPF to construct an accurate picture of the network. Example 8-4 shows the LSAs in area 3 (Figures 8-6 and 8-7) via show commands. Example 8-4

LSA Types 1 and 2 in Area 3

! R3’s LSDB is shown, with type 1 LSAs listed as “Router Link States” and ! type 2 LSAs as “Net Link States.” The command output shows a section for each LSA ! type, in sequential order. R3# show ip ospf database OSPF Router with ID (3.3.3.3) (Process ID 1) Router Link States (Area 3)

continues

274

Chapter 8: OSPF

Example 8-4

LSA Types 1 and 2 in Area 3 (Continued)

Link ID

ADV Router

Age

Seq#

1.1.1.1

1.1.1.1

1203

0x80000025 0x0072C3 2

Checksum Link count

3.3.3.3

3.3.3.3

779

0x80000027 0x003FB0 3

10.3.3.33

10.3.3.33

899

0x80000020 0x002929 2

Net Link States (Area 3) Link ID

ADV Router

Age

Seq#

10.3.1.3

3.3.3.3

1290

0x8000001F 0x00249E

Checksum

! Lines omitted for brevity ! Next, the specific LSA’s link ID is included in the show command, listing detail ! for the one LSA type 2 inside area 3. Note that the “Link ID” is the DR’s ! interface address on the subnet. The network keyword refers to the network LSAs (type 2 LSAs). R3# show ip ospf database network 10.3.1.3 OSPF Router with ID (3.3.3.3) (Process ID 1) Net Link States (Area 3) Routing Bit Set on this LSA LS age: 1304 Options: (No TOS-capability, DC) LS Type: Network Links Link State ID: 10.3.1.3 (address of Designated Router) Advertising Router: 3.3.3.3 LS Seq Number: 8000001F Checksum: 0x249E Length: 32 Network Mask: /23 Attached Router: 3.3.3.3 Attached Router: 10.3.3.33 ! Next, the type 1 LSA for R3 is listed. The link ID is the RID of R3. Note that ! the LSA includes reference to each stub and transit link connected to R3. The router ! keyword refers to the router LSAs (type 1 LSAs). R3# show ip ospf database router 3.3.3.3 OSPF Router with ID (3.3.3.3) (Process ID 1) Router Link States (Area 3) LS age: 804 Options: (No TOS-capability, DC) LS Type: Router Links Link State ID: 3.3.3.3 Advertising Router: 3.3.3.3 LS Seq Number: 80000027 Checksum: 0x3FB0 Length: 60 Number of Links: 3 Link connected to: another Router (point-to-point) (Link ID) Neighboring Router ID: 1.1.1.1 (Link Data) Router Interface address: 10.3.13.3 Number of TOS metrics: 0 TOS 0 Metrics: 64

OSPF Design and LSAs

Example 8-4

275

LSA Types 1 and 2 in Area 3 (Continued) Link connected to: a Stub Network (Link ID) Network/subnet number: 10.3.13.0 (Link Data) Network Mask: 255.255.255.0 Number of TOS metrics: 0 TOS 0 Metrics: 64

! Note that R3’s LSA refers to a transit network next, based on its DR RID – ! these lines allow OSPF to know that this router (R3) connects to the transit ! network whose type 2 LSA has LSID 10.3.1.3. Link connected to: a Transit Network (Link ID) Designated Router address: 10.3.1.3 (Link Data) Router Interface address: 10.3.1.3 Number of TOS metrics: 0 TOS 0 Metrics: 10 ! Below, the routes from R3 and R1 to 10.3.2.0/23 are shown. Note the cost values ! for each reflect the cumulative costs of the outgoing interfaces used to reach ! the subnet—for instance, R3’s cost is the sum of its outgoing interface cost ! (10) plus R33’s outgoing interface cost (1). R1’s cost is based on three outgoing ! links: R1 (cost 64), R3 (cost 10), and R33 (cost 1), for a total of 75. Also ! note that the time listed in the route is the time since this LSA first arrived ! at the router, even if the LSA has been refreshed due to the LSRefresh interval. R3# show ip route ospf 1 | include 10.3.2.0 O

10.3.2.0/23 [110/11] via 10.3.1.33, 17:08:33, Ethernet0/0

R1# show ip route ospf | include 10.3.2.0 O

10.3.2.0/23 [110/75] via 10.3.13.3, 17:10:15, Serial0/0.3

The show ip ospf database command lists the LSAs in that router’s LSDB, with LSA type 1 LSAs (router LSAs) first, then type 2 (network link states), continuing sequentially through the LSA types. Also note that the LSDB for area 3 should be identical on R33, R3, and R1. However, on R1, the show ip ospf database command lists all of R1’s LSDB entries, including LSAs from other areas, so using an internal router to look at the LSDB may be the best place to begin troubleshooting a problem. Also note the costs for the routes on R3 and R1 at the end of the example—the SPF algorithm simply added the outgoing costs along the routes, from each router’s perspective. NOTE To signify a network that is down, the appropriate type 1 or 2 LSA is changed to show a metric of 16,777,215 (224 – 1), which is considered to be an infinite metric to OSPF.

LSA Type 3 and Inter-Area Costs ABRs do not forward type 1 and 2 LSAs from one area to another. Instead, ABRs advertise type 3 LSAs into one area in order to represent subnets described in both the type 1 and 2 LSAs in another area. Each type 3 summary LSA describes a simple vector—the subnet, mask, and the ABR’s cost to reach that subnet, as shown in Figure 8-8.

276

Chapter 8: OSPF

Figure 8-8

Representation of Area 3 Subnets as Type 3 LSAs in Area 0 Add: My cost to reach R1 (cost 1) to Type 3 LSA’s costs

Area 0 R2’s Three Type 3 LSAs: Subnet 10.3.2.0/23, Cost 75 Subnet 10.3.0.0/23, Cost 74 Subnet 10.3.13.0/24, Cost 64

R1

S2 Routing Table: 10.3.2.0/23: Cost 76 10.3.0.0/23: Cost 75 10.3.13.0/24: Cost 65

Example 8-5 focuses on the three subnets inside area 3, looking at the type 3 summary LSAs created for those subnets by ABR R1. Note that the example shows commands on S2; S2 has identical area 0 LSDB entries as compared with R1. Example 8-5

LSA Type 3 Created by R1 for Area 3’s Subnets

! S2, internal to area 0, does not have the type 1 and 2 LSAs seen by R3 back in ! Example 8-4. However, type 3 LSAs (listed as “Summary Net Links”) show all ! three subnets inside area 3. R1 is listed as the advertising router because it ! created the type 3 LSAs. S2# show ip ospf database ! Lines omitted for brevity Summary Net Link States (Area 0) Link ID

ADV Router

Age

Seq#

10.3.0.0

1.1.1.1

257

0x80000001 0x00A63C

Checksum

10.3.2.0

1.1.1.1

257

0x80000001 0x009A45

10.3.13.0

1.1.1.1

261

0x80000021 0x007747

! Lines omitted for brevity ! Below, note that the summary keyword is used to view type 3 LSAs. The metric ! reflects R1’s cost to reach the subnet inside area 3. S2# show ip ospf database summary 10.3.0.0 OSPF Router with ID (8.8.8.8) (Process ID 1) Summary Net Link States (Area 0) Routing Bit Set on this LSA LS age: 341 Options: (No TOS-capability, DC, Upward) LS Type: Summary Links(Network) Link State ID: 10.3.0.0 (summary Network Number) Advertising Router: 1.1.1.1 LS Seq Number: 80000001 Checksum: 0xA63C Length: 28 Network Mask: /23 TOS: 0

Metric: 74

! Next, S2’s routes to all three subnets are listed. S2 calculates its cost ! based on its cost to reach R1, plus the cost listed in the type 3 LSA. For

OSPF Design and LSAs

Example 8-5

277

LSA Type 3 Created by R1 for Area 3’s Subnets (Continued)

! example, the cost (above) in the type 3 LSA for 10.3.0.0/23 is 74; S2 adds ! that to S2’s cost to reach ABR R1 (cost 1), for a metric of 75. S2# show ip route ospf | include 10.3 O IA

10.3.13.0/24 [110/65] via 10.1.1.1, 00:16:04, Vlan1

O IA

10.3.0.0/23 [110/75] via 10.1.1.1, 00:05:08, Vlan1

O IA

10.3.2.0/23 [110/76] via 10.1.1.1, 00:05:12, Vlan1

! Next, S2’s cost to reach RID 1.1.1.1 is listed as cost 1. S2# show ip ospf border-routers OSPF Process 1 internal Routing Table Codes: i—Intra-area route, I—Inter-area route i 1.1.1.1 [1] via 10.1.1.1, Vlan1, ABR, Area 0, SPF 18 i 2.2.2.2 [1] via 10.1.1.2, Vlan1, ABR, Area 0, SPF 18 i 7.7.7.7 [1] via 10.1.1.3, Vlan1, ASBR, Area 0, SPF 18 ! Below, the show ip ospf statistics command lists the number of SPF calculations. R1# show ip ospf stat OSPF process ID 1 -----------------------------------------Area 0: SPF algorithm executed 6 times Area 3: SPF algorithm executed 15 times Area 4: SPF algorithm executed 6 times Area 5: SPF algorithm executed 5 times ! Lines omitted for brevity

Example 8-5 shows how S2 calculated its cost to the area 3 subnets. Routers calculate the cost for a route to a subnet defined in a type 3 LSA by adding the following items: 1.

The calculated cost to reach the ABR that created and advertised the type 3 LSA.

2.

The cost as listed in the type 3 LSA.

You can see the cost of the type 3 LSA with the show ip ospf database summary link-id command, and the cost to reach the advertising ABR with the show ip ospf border-routers command, as shown in Example 8-5. The beauty of this two-step cost calculation process is that it allows a significant reduction in the number of SPF calculations. When a type 1 or 2 LSA changes in some way that affects the underlying routes—for instance, a link failure—each router in the area runs SPF, but routers inside other areas do not. For instance, if R3’s E0/0 is shut down, all three routers in area 3 run SPF inside that area, and the counter for area 3 in the show ip ospf statistics command increments. However, routers not inside area 0 do not run SPF, even though they update their routing tables—a process called a partial run, partial SPF, or partial calculation. For example, imagine that R3’s LAN interface fails. R33 then updates its type 2 LSA, listing a metric of 16,777,215. R1 in turn updates its type 3 LSA for 10.3.0.0/23, flooding that throughout

278

Chapter 8: OSPF

area 0. The next step shows the computational savings: S2, using the two-step calculation, simply adds its cost to R1 (still 1) to 16,777,215, finds the number out of range, and removes the route from the IP routing table. S2 did not have to actually run the SPF algorithm to discover a new SPF tree. Of particular importance is that partial calculations happen without any route summarization. With OSPF, route summarization does help reduce the overall number of routes that require SPF calculations, but route summarization is not required for partial calculations to occur. Removing Routes Advertised by Type 3 LSAs When a router wants to remove a route advertised by a type 3 LSA from the LSDBs of its neighbors, it could simply remove that route from its LSDB and stop advertising it. The trouble with that approach is that the route might stick around for a while in other routers’ LSDBs. Clearly, it is better to actively remove the failed route instead. As a result, the router that was advertising the failed route sets the route’s age to the Maxage, as described in RFC 2328, and refloods it throughout the routing domain. This removes the route as quickly as possible from the domain, rather than waiting for it to age out slowly. LSA Types 4 and 5, and External Route Types 1 and 2 OSPF allows for two types of external routes, aptly named types 1 and 2. The type determines whether only the external metric is considered by SPF when picking the best routes (external type 2, or E2), or whether both the external and internal metrics are added together to compute the metric (external type 1, or E1). When an ASBR injects an E2 route, it creates a type 5 LSA for the subnet. The LSA lists the metric. The ASBR then floods the type 5 LSA throughout all areas. The other routers simply use the metric listed in the LSA; no need exists to add any cost on any links internal to the OSPF domain. When an ABR then floods the type 5 LSA into another area, the ABR creates a type 4 LSA, listing the ABR’s metric to reach the ASBR that created the type 5 LSA. Other routers calculate their costs to reach E1 routes in a manner similar to how metrics for LSA type 3 routes are calculated— by calculating the cost to reach the ASBR, and then adding the cost listed in the type 5 LSA. Figure 8-9 outlines the mechanics of how the LSAs are propagated, and how the metrics are calculated. Figure 8-9

LSA Types 4 and 5 Propagation and the Effect on Type 1 External Routes Area 3

3

Internal Router

ABR RID 1.1.1.1

R3

R1

• Cost to E2 Route: Metric in LSA Type 5 • Cost to E1 Route: Add: My Cost to ABR R1, Plus the LSA Type 4’s Cost, Plus the LSA Type 5’s Cost

Area 0

ASBR RID 7.7.7.7 S1

2

• Create/Flood Type 4: List R1’s Metric to Reach the ASBR • Flood Type 5’s, Metric Unchanged

Note: Arrows Show Propagation of LSAs.

1 • Create/Flood Type 5’s, Metric 20

Externals: 192.168.1.0/24 (E1) 192.168.2.0/24 (E2)

OSPF Design and LSAs

279

E1 routes by definition include the cost as assigned when the ASBR injected the route into OSPF, plus any cost inside the OSPF domain. To calculate the cost for the E1 route, a router inside a different area than the ASBR must use two steps to calculate the internal cost, and a third step to add the external cost. For example, when R3, internal to area 3, calculates the cost to reach 192.168.1.0/24 (an E1 route), R3 adds the following: ■

R3’s calculated area 3 cost to reach ABR R1 (RID 1.1.1.1).

■

R1’s cost to reach the ASBR that advertised the route (S2, RID 7.7.7.7). R1 announces this cost in the LSA type 4 that describes R1’s cost to reach ASBR 7.7.7.7.

■

The external metric for the route, as listed in the type 5 LSA created by the ASBR.

Example 8-6 shows the components of the metrics and LSAs for two external routes: 192.168.1.0/24 E1 with metric 20, and 192.168.2.0/24 E2, also with metric 20. Example 8-6

Calculating the Metric for External Types 1 and 2

! R3 has learned the two LSA type 5s. R3# show ip ospf database | begin Type-5 Type-5 AS External Link States Link ID

ADV Router

Age

Seq#

192.168.1.0

7.7.7.7

1916

0x8000002B 0x0080EF 0

Checksum Tag

192.168.2.0

7.7.7.7

1916

0x80000028 0x00FEF2 0

! Next, the detail for E2 192.168.2.0 is listed, with “metric type” referring ! to the external route type E2. (192.168.1.0, not shown, is type 1.) R3# show ip ospf database external 192.168.2.0 OSPF Router with ID (3.3.3.3) (Process ID 1) Type-5 AS External Link States Routing Bit Set on this LSA LS age: 1969 Options: (No TOS-capability, DC) LS Type: AS External Link Link State ID: 192.168.2.0 (External Network Number) Advertising Router: 7.7.7.7 LS Seq Number: 80000028 Checksum: 0xFEF2 Length: 36 Network Mask: /24 Metric Type: 2 (Larger than any link state path) TOS: 0 Metric: 20 Forward Address: 0.0.0.0 External Route Tag: 0 ! Next, R1’s advertised cost of 1 between itself and the ASBR is listed. Note ! that S1’s RID (7.7.7.7) is listed, with the ABR that forwarded the LSA into

continues

280

Chapter 8: OSPF

Example 8-6

Calculating the Metric for External Types 1 and 2 (Continued)

! area 3, R1 (RID 1.1.1.1) also listed. R3# show ip ospf database asbr-summary OSPF Router with ID (3.3.3.3) (Process ID 1) Summary ASB Link States (Area 3) Routing Bit Set on this LSA LS age: 923 Options: (No TOS-capability, DC, Upward) LS Type: Summary Links(AS Boundary Router) Link State ID: 7.7.7.7 (AS Boundary Router address) Advertising Router: 1.1.1.1 LS Seq Number: 8000000A Checksum: 0x12FF Length: 28 Network Mask: /0 TOS: 0

Metric: 1

! Below, R3’s calculated cost to R1 (64) and then to S2 (7.7.7.7) are listed. Note ! that the total of 65 is the cost 64 to reach the ABR, plus the cost 1 for the ! ABR to reach the ASBR. R3# show ip ospf border-routers OSPF Process 1 internal Routing Table Codes: i—Intra-area route, I—Inter-area route i 1.1.1.1 [64] via 10.3.13.1, Serial0/0.1, ABR, Area 3, SPF 30 I 7.7.7.7 [65] via 10.3.13.1, Serial0/0.1, ASBR, Area 3, SPF 30 ! Below, each route is noted as E1 or E2, with the E1 route’s metric including ! the external cost (20), plus cost to reach the ASBR (65). R3# show ip route | include 192.168 O E1 192.168.1.0/24 [110/85] via 10.3.13.1, 00:50:34, Serial0/0.1 O E2 192.168.2.0/24 [110/20] via 10.3.13.1, 00:50:34, Serial0/0.1

OSPF Design in Light of LSA Types OSPF’s main design trade-offs consist of choosing links for particular areas, with the goal of speeding convergence, reducing memory and computing resources, and keeping routing tables small through route summarization. For instance, by using a larger number of areas, and the implied conversion of dense types 1 and 2 LSAs into sparser type 3 LSAs, the OSPF LSDBs can be made smaller. Also, link flaps in one area require SPF calculations only in that area, due to the partial calculation feature. Additionally, ABRs and ASBRs can be configured to summarize routes, reducing the number of type 3 LSAs introduced into other areas as well. (Route summarization is covered in Chapter 9.) The OSPF design goals to reduce convergence time, reduce overhead processing, and improve network stability can be reached using the core OSPF protocols and features covered so far. Another key OSPF design tool, stubby areas, will be covered next.

OSPF Design and LSAs

281

NOTE Before moving on, a comment is in order about the relative use of the word “summary” in OSPF. The typical uses within OSPF include the following: ■ Type 3 LSAs are called summary LSAs in the OSPF RFCs. ■

Type 5 and 7 external LSAs are sometimes called summary LSAs, because the LSAs cannot represent detailed topology information.

■

The term LSA summary refers to the LSA headers that summarize LSAs and are sent inside DD packets.

■

The term summary can also be used to refer to summary routes created with the area range and summary-address commands.

Stubby Areas OSPF can further reduce overhead by treating each area with one of several variations of rules, based on a concept called a stubby area. Stubby areas take advantage of the fact that to reach subnets in other areas, routers in an area must forward the packets to some ABR. Without stubby areas, ABRs must advertise all the subnets into the area, so that the routers know about the subnets. With stubby areas, ABRs quit advertising type 5 (external) LSAs into the stubby area, but instead ABRs create and advertise default routes into the stubby area. As a result, internal routers use default routing to forward packets to the ABR anyway. However, the internal routers now have sparser LSDBs inside the area. The classic case for a stubby area is an area with one ABR, but stubby areas can work well for areas with multiple ABRs as well. For example, the only way out of area 3 in Figure 8-6 is through the only ABR, R1. So, R1 could advertise a default route into area 3 instead of advertising any external type 5 LSAs. Also in Figure 8-6, area 5 has two ABRs. If area 5 were a stubby area, both ABRs would inject default routes into the area. This configuration would work, but it may result in suboptimal routing. OSPF defines several different types of stubby areas. By definition, all stubby areas stop type 5 (external) LSAs from being injected into them by the ABRs. However, depending on the variation, a stubby area may also prevent type 3 LSAs from being injected. The other variation includes whether a router inside the stubby area can redistribute routes into OSPF, thereby injecting an external route. Table 8-5 lists the variations on stubby areas, and their names. Note in Table 8-5 that all four stub area types stop type 5 LSAs from entering the area. When the name includes “totally,” type 3 LSAs are also not passed into the area, significantly reducing the size of the LSDB. If the name includes “NSSA,” it means that external routes can be redistributed into OSPF by routers inside the stubby area; note that the LSAs for these external routes would be type 7.

282

Chapter 8: OSPF

OSPF Stubby Area Types

Table 8-5

Area Type

Stops Injection of Type 5 LSAs?

Stops Injection of Type 3 LSAs?

Allows Creation of Type 7 LSAs Inside the Area?

Stub

Yes

No

No

Totally stubby

Yes

Yes

No

Not-so-stubby area (NSSA)

Yes

No

Yes

Totally NSSA

Yes

Yes

Yes

Configuring a stub area is pretty simple—all routers in the area need the same stub settings, as configured in the area stub command. Table 8-6 lists the options. Stub Area Configuration Options

Table 8-6

Stub Type

Router OSPF Subcommand

NSSA

area area-id nssa

Totally NSSA

area area-id nssa no-summary

Stub

area area-id stub

Totally stubby

area area-id stub no-summary

Example 8-7, based on Figure 8-6, shows the results of the following configuration: ■

Area 3 is configured as a totally NSSA area.

■

R3 will inject an external route to 192.168.21.0/24 as a type 7 LSA.

■

Area 4 is configured as a totally stubby area.

■

Area 5 is configured as simply stubby.

Example 8-7

Stub Area Example

! R3, in a totally NSSA area, knows intra-area routes (denoted with an “IA” ! near the front of the output line from show ip route), but the only ! interarea route is the default route created and sent by R1, the ABR. R3# show ip route ospf 10.0.0.0/8 is variably subnetted, 3 subnets, 2 masks O

10.3.2.0/23 [110/11] via 10.3.1.33, 00:00:00, Ethernet0/0

O*IA 0.0.0.0/0 [110/65] via 10.3.13.1, 00:00:00, Serial0/0.1 ! Still on R3, the LSA type 3 summary, created by ABR R1, is shown first. ! Next, the External NSSA LSA type 7 LSA created by R3 is listed.

OSPF Design and LSAs

Example 8-7

283

Stub Area Example (Continued)

R3# show ip ospf database | begin Summary Summary Net Link States (Area 3) Link ID

ADV Router

Age

Seq#

0.0.0.0

1.1.1.1

704

0x80000004 0x00151A

Checksum

Type-7 AS External Link States (Area 3) Link ID

ADV Router

Age

Seq#

192.168.21.0

3.3.3.3

17

0x80000003 0x00C12B 0

Checksum Tag

! R1, because it is attached to area 3, also has the R3-generated NSSA external ! LSA. Note the advertising router is R3, and it is an E2 external route. R1# show ip ospf database nssa-external OSPF Router with ID (1.1.1.1) (Process ID 1) Type-7 AS External Link States (Area 3) Routing Bit Set on this LSA LS age: 188 Options: (No TOS-capability, Type 7/5 translation, DC) LS Type: AS External Link Link State ID: 192.168.21.0 (External Network Number ) Advertising Router: 3.3.3.3 LS Seq Number: 80000003 Checksum: 0xC12B Length: 36 Network Mask: /24 Metric Type: 2 (Larger than any link state path) TOS: 0 Metric: 20 Forward Address: 10.3.13.3 External Route Tag: 0 ! Below, the same command on R2, not in area 3, shows no type 7 LSAs. ABRs ! convert type 7 LSAs to type 5 LSAs before forwarding them into another area. R2# show ip ospf database nssa-external OSPF Router with ID (2.2.2.2) (Process ID 2) ! Next, R2 does have a type 5 LSA for the subnet; R1 converts the type 7 to a type ! 5 before flooding it into other areas. R2# show ip ospf database | begin Type-5 Type-5 AS External Link States Link ID

ADV Router

Age

Seq#

192.168.1.0

7.7.7.7

521

0x80000050 0x003615 0

Checksum Tag

192.168.2.0

7.7.7.7

521

0x8000004D 0x00B418 0

192.168.21.0

1.1.1.1

1778

0x80000019 0x006682 0

continues

284

Chapter 8: OSPF

Example 8-7

Stub Area Example (Continued)

! Below, R4 is in a totally stubby area, with only one inter-area route. R4# show ip route ospf O*IA 0.0.0.0/0 [110/1563] via 10.4.14.1, 00:11:59, Serial0/0.1 ! R5, in a stubby area, has several inter-area routes, but none of the ! external routes (e.g. 192.168.1.0). R5’s default points to R2. R5# show ip route ospf 10.0.0.0/8 is variably subnetted, 7 subnets, 3 masks O IA

10.3.13.0/24 [110/115] via 10.5.25.2, 13:45:49, Serial0.2

O IA

10.3.0.0/23 [110/125] via 10.5.25.2, 13:37:55, Serial0.2

O IA

10.1.1.0/24 [110/51] via 10.5.25.2, 13:45:49, Serial0.2

O IA

10.4.0.0/16 [110/1613] via 10.5.25.2, 13:45:49, Serial0.2

O*IA 0.0.0.0/0 [110/51] via 10.5.25.2, 13:45:49, Serial0.2 ! Below, R5’s costs on its two interfaces to R1 and R2 are highlighted. Note that ! the default route’s metric (51) comes from the 50 below, plus an advertised ! cost of 1 in the summary (type 3) for default 0.0.0.0/0 generated by R2. R5 ! simply chose to use the default route with the lower metric. R5# sh ip ospf int brief Interface

PID

Area

IP Address/Mask

Cost

State Nbrs F/C

Se0.1

1

5

10.5.15.5/24

64

P2P

1/1

Se0.2

1

5

10.5.25.5/24

50

P2P

1/1

Et0

1

5

10.5.1.5/24

10

DR

0/0

! Next, R2 changes the cost of its advertised summary from 1 to 15. R2# conf t Enter configuration commands, one per line.

End with CNTL/Z.

R2(config)# router ospf 2 R2(config-router)# area 5 default-cost 15 ! Below, R5’s metrics to both R1’s and R2’s default routes tie, ! so both are now in the routing table. R5# show ip route ospf ! lines omitted for brevity O*IA 0.0.0.0/0 [110/65] via 10.5.25.2, 00:00:44, Serial0.2 [110/65] via 10.5.15.1, 00:00:44, Serial0.1

The legend in the top of the output of a show ip route command lists several identifiers that pertain to OSPF. For example, the acronym “IA” refers to interarea OSPF routes, E1 refers to external type 1 routes, and E2 refers to external type 2 routes.

Graceful Restart In steady-state operation, OSPF can react to changes in the routing domain and reconverge quickly. This is one of OSPF’s strengths as an IGP. However, what happens when something goes really wrong is just as important as how things work under relatively stable conditions.

OSPF Design and LSAs

285

One of those “really wrong” things that sometimes happens is that a router requires a restart to its OSPF software process. To prevent various routing problems, including loops, that can take place when an OSPF router suddenly goes away while its OSPF software is restarting is the graceful restart process documented in RFC 3623. Cisco implemented its own version of graceful restart in Cisco IOS prior to RFC 3623; as a result, Cisco IOS supports both versions. Graceful restart is also known as nonstop forwarding (NSF) in RFC 3623 because of the way it works. Graceful restart takes advantage of the fact that modern router architectures use separate routing and forwarding planes. It is possible to continue forwarding without loops while the routing process restarts, assuming the following conditions are true: ■

The router whose OSPF process is restarting must notify its neighbors that the restart is going to take place by sending a “grace LSA.”

■

The LSA database remains stable during the restart.

■

All of the neighbors support, and are configured for, graceful restart.

■

The restart takes place within a specific “grace period.”

■

During restart, the neighboring fully adjacent routers must operate in “helper mode.”

In Cisco IOS, CEF handles forwarding during graceful restart while OSPF rebuilds the RIB tables, provided that the preceding conditions are met. Both Cisco and IETF NSF support are enabled by default in Cisco IOS, beginning with version 12.4(6)T. Disabling it requires a routing process command for each NSF version, nsf [cisco | ietf] helper disable.

OSPF Path Choices That Do Not Use Cost Under most circumstances, when an OSPF router runs the SPF algorithm and finds more than one possible route to reach a particular subnet, the router chooses the route with the least cost. However, OSPF does consider a few conditions other than cost when making this best-path decision. This short section explains the remaining factors that impact which route, or path, is considered best by the SPF algorithm. Choosing the Best Type of Path As mentioned earlier, some routes are considered to be intra-area routes, some are interarea routes, and two are types of external routes (E1 and E2). It is possible for a router to find multiple routes to reach a given subnet where the type of route (intra-area, interarea, E1, or E2) is different. In these cases, RFC 2328 specifies that the router should ignore the costs and instead chooses the best route based on the following order of preference: 1.

Intra-area routes

2.

Interarea routes

286

Chapter 8: OSPF 3.

E1 routes

4.

E2 routes

For example, if a router using OSPF finds one intra-area route for subnet 1 and one interarea route to reach that same subnet, the router ignores the costs and simply chooses the intra-area route. Similarly, if a router finds one interarea route, one E1 route, and one E2 route to reach the same subnet, that router chooses the interarea route, again regardless of the cost for each route. Best-Path Side Effects of ABR Loop Prevention The other item that affects OSPF best-path selection relates to some OSPF loop-avoidance features. Inside an area, OSPF uses Link State logic, but between areas OSPF acts much like a Distance Vector (DV) protocol in some regard. For example, the advertisement of a Type 3 LSA from one area to another hides the topology in the original area from the second area, just listing a destination subnet, metric (cost), and the ABR through which the subnet can be reached—all DV concepts. OSPF does not use all the traditional DV loop avoidance features, but it does use some of the same underlying concepts, including Split Horizon. In OSPF’s case, it applies Split Horizon for several types of LSAs so that an LSA is not advertised into one nonbackbone area and then advertised back into the backbone area. Figure 8-10 shows an example in which ABR1 and ABR2 both advertise Type 3 LSAs into area 1, but then they both choose to not forward a Type 3 LSA back into area 0. Figure 8-10

Split Horizon per Area with OSPF Subnet 1

Area 2

Cost 1

ABR3

Area 0

Cost 1

Cost 100

Type 3 LSAs

ABR1

ABR2 Cost 1

Area 1

Cost 1

R1

Cost 1

R2

OSPF Design and LSAs

287

The figure shows the propagation of some of the LSAs for subnet 1 in this figure but not all. ABR3 generates a T3 LSA for subnet 1 and floods that LSA within area 0. ABR1 floods a T3 LSA for subnet 1 into area 1; however, when ABR2 gets that T3 LSA from ABR1, ABR2 does not flood a T3 LSA back into area 0. (To reduce clutter, the figure does not include arrowed lines for the opposite direction, in which ABR2 floods a T3 LSA into area 1, and then ABR1 chooses not to flood a T3 LSA back into area 0.) More generically speaking, an ABR can learn about summary LSAs from other ABRs, inside the nonbackbone area, but the ABR will not then advertise another LSA back into area 0 for that subnet. Although interesting, none of these facts impacts OSPF path selection. The second part of ABR loop prevention is the part that impacts path selection, as follows: ABRs ignore LSAs created by other ABRs, when learned through a nonbackbone area, when calculating least-cost paths. This prevents an ABR from choosing a path that goes into one nonbackbone area and then back into area 0 through some other ABR. For example, without this rule, in the internetwork of Figure 8-11, router ABR2 would calculate a cost 3 path to subnet 1: from ABR2 to ABR1 inside area 1 and then from ABR1 to ABR3 in area 0. ABR2 would also calculate a cost 101 path to subnet 1, going from ABR2 through area 0 to ABR3. Clearly, the first of these two paths, with cost 3, is the least-cost path. However, ABRs use this additional loop-prevention rule, meaning that ABR2 ignores the T3 LSA advertised by ABR1 for subnet 1. This behavior prevents ABR2 from choosing the path through ABR2, so in actual practice, ABR2 would find only one possible path to subnet 1: the path directly from ABR2 to ABR3. Figure 8-11

Effect of ABR2 Ignoring Path to Subnet 1 Through Area 1 Subnet 1

Area 2 Cost 1

ABR3 Cost 3 path

Area 0

Cost 101 path Cost 1

ABR1

Area 1

Cost 1

R1

Cost 100

Cost 1

ABR2

Cost 1

R2

288

Chapter 8: OSPF

It is important to notice that the link between ABR1 and ABR2 is squarely inside nonbackbone area 1. If this link were in area 0, ABR2 would pick the best route to reach ABR3 as being ABR2— ABR1—ABR3, choosing the lower-cost route. This loop-prevention rule has some even more interesting side effects for internal routers. Again in Figure 8-10, consider the routes calculated by internal router R2 to reach subnet 1. R2 learns a T3 LSA for subnet 1 from ABR1, with the cost listed as 2. To calculate the total cost for using ABR1 to reach subnet 1, R2 adds its cost to reach ABR1 (cost 2), totaling cost 4. Likewise, R2 learns a T3 LSA for subnet 1 from ABR2, with cost 101. R2 calculates its cost to reach ABR2 (cost 1) and adds that to 101 to arrive at cost 102 for this alternative route. As a result, R1 picks the route through ABR1 as the best route. However, the story gets even more interesting with the topology in Figure 8-10. R2’s next-hop router for the R2—ABR2—ABR1—ABR3 path is ABR2. So, R2 forwards packets destined to subnet 1 to ABR2 next. However, as noted just a few paragraphs ago, ABR2’s route to reach subnet 1 points directly to ABR3. As a result, packets sent by R2, destined to subnet 1, actually take the path from R2 —ABR2—ABR3. As you can see, these decisions can result in arguably suboptimal routes, and even asymmetric routes, as would be the case in this particular example.

OSPF Configuration This section covers the core OSPF configuration commands, along with the OSPF configuration topics not already covered previously in the chapter. (If you happened to skip the earlier parts of this chapter, planning to review OSPF configuration, make sure to go back and look at the earlier examples in the chapter. These examples cover OSPF stubby area configuration, OSPF network types, plus OSPF neighbor and priority commands.) Example 8-8 shows configuration for the routers in Figure 8-6, with the following design goals in mind: ■

Proving that OSPF PIDs do not have to match on separate routers

■

Using the network command to match interfaces, thereby triggering neighbor discovery inside network 10.0.0.0

■

Configuring S1’s RID as 7.7.7.7

OSPF Configuration

■

Setting priorities on the backbone LAN to favor S1 and S2 to become the DR/BDR

■

Configuring a minimal dead interval of 1 second, with hello multiplier of 4, yielding a 250-ms hello interval on the backbone LAN

Example 8-8

289

OSPF Configuration Basics and OSPF Costs

! R1 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! ! R1 has been configured for a (minimal) 1-second dead interval, and 1/4-second ! (250 ms) hello interval based on 4 Hellos per 1-second dead interval. interfac e FastEthernet0/ 0 i p a d d r e s s 1 0 . 1 . 1 . 1 2 5 5 .2 5 5 . 2 5 5 . 0 ip ospf dead-interval minimal hello-multiplier 4 ! R1 uses the same stub area configuration as in Example 8-7, with network ! commands matching based on the first two octets. Note that the network commands ! place each interface into the correct area. router ospf 1 area 3 ns sa no-summary area 4 stub no-summary area 5 stub n e tw o r k 1 0 . 1 . 0 . 0 0 . 0 . 2 5 5 . 2 5 5 a r e a 0 network 10.3.0.0 0.0.255.255 area 3 n e t w o r k 1 0 . 4 . 0 . 0 0 . 0 . 2 5 5 . 2 5 5 a r ea 4 n e t w o r k 1 0 . 5 . 0 . 0 0 . 0 . 25 5 . 2 5 5 a r e a 5 ! R2 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! ! The R2 configuration also uses the Fast Hello feature, otherwise it ! would not match hello and dead intervals with R1. interface FastEthernet0/0 ip address 10.1.1.2 255.255.255.0 i p o s p f d e a d - i n t e r v a l m in i m a l h e l l o - m u l t i p l i e r 4 ! Below, R2 uses a different PID than R1, but the PID is only used locally. ! R1 and R2 will become neighbors. Also, all routers in a stubby area must be ! configured to be that type of stubby area; R2 does that for area 5 below. router ospf 2 area 5 stub network 10.1.0.0 0.0.255.255 area 0 n e t w o r k 1 0 . 5 .2 5 . 2 0 . 0 . 0 . 0 a r e a 5 ! R3 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! !Note that R3’s area 2 nssa no-summary command must match the settings ! on R1. Likewise, below, R4’s stub settings must match R1’s settings for area 4. . rout er ospf 1 area 3 nssa no-s ummary n e t w o r k 1 0 . 0 .0 . 0 0 . 2 5 5 . 2 5 5 . 2 5 5 a r e a 3 ! R4 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! router ospf 1 are a 4 stub no-sum mary network 10.0.0.0 0.255.255.255 area 4

continues

290

Chapter 8: OSPF

Example 8-8

OSPF Configuration Basics and OSPF Costs (Continued)

! S1 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! ! S1 matches hello and dead intervals on the LAN. Also, it sets its OSPF ! priority to 255, the maximum value, hoping to become the DR. interfac e Vlan1 i p a d d r e ss 1 0 . 1 . 1 . 3 2 5 5 . 2 5 5 . 2 5 5 . 0 i p o s p f d e a d - i n t e r v a l m i ni m a l h e l l o - m u l t i p l i e r 4 i p ospf priority 255 ! Below, S1 sets its RID manually, removing any reliance on an interface address. route r ospf 1 r o u t e r - i d 7. 7 . 7 . 7 n e t w o r k 1 0 . 1 . 0 . 0 0. 0 . 2 5 5 . 2 5 5 a r e a 0 ! S2 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! ! Below, S2 also matches timers, and sets its priority to 1 less than S1, hoping ! to be the BDR. int erface Vlan1 ip address 10.1.1.4 255.255.255.0 i p o s p f d e a d - i n t e r v a l m i ni m a l h e l l o - m u l t i p l i e r 4 ip ospf pri ority 254 ! router ospf 1 network 10.0.0.0 0.255.255.255 area 0

Note that R3 and R4 do not need the no-summary option on the area command; this parameter is only needed at the ABR, in this case R1. The parameters are shown here to stress the variations of stubby areas. OSPF Costs and Clearing the OSPF Process Example 8-9 highlights a few details about clearing (restarting) the OSPF process, and looks at changes to OSPF costs. This example shows the following sequence: 1.

R3’s OSPF process is cleared, causing all neighbors to fail and restart.

2.

R3’s log-adjacency-changes detail configuration command (under router ospf) causes more detailed neighbor state change messages to appear.

3.

R5 has tuned its cost settings with the ip ospf cost 50 interface subcommand under S0.2 in order to prefer R2 over R1 for reaching the core.

4.

R2 is configured to use a new reference bandwidth, changing its cost calculation per interface.

Example 8-9

Changing RIDs, Clearing OSPF, and Cost Settings

R3# clear ip ospf process Reset ALL OSPF processes? [no]: y ! Above, all OSPF processes are cleared on R3. R3 has the log-adjacency-changes ! detail command configured, so that a message is generated at each state ! change, as shown below for neighbor R33 (RID 192.168.1.1). (Messages for ! other routers are omitted.)

OSPF Configuration

291

Changing RIDs, Clearing OSPF, and Cost Settings (Continued)

Example 8-9

00:02:46: %OSPF-5-ADJCHG: Process 1, Nbr 192.168.1.1 on Ethernet0/0 from FULL to DOWN, Neighbor Down: Interface down or detached 00:02:53: %OSPF-5-ADJCHG: Process 1, Nbr 192.168.1.1 on Ethernet0/0 from DOWN to INIT, Received Hello 00:02:53: %OSPF-5-ADJCHG: Process 1, Nbr 192.168.1.1 on Ethernet0/0 from INIT to 2WAY, 2-Way Received 00:02:53: %OSPF-5-ADJCHG: Process 1, Nbr 192.168.1.1 on Ethernet0/0 from 2WAY to EXSTART, AdjOK? 00:02:53: %OSPF-5-ADJCHG: Process 1, Nbr 192.168.1.1 on Ethernet0/0 from EXSTART to EXCHANGE, Negotiation Done 00:02:53: %OSPF-5-ADJCHG: Process 1, Nbr 192.168.1.1 on Ethernet0/0 from EXCHANGE to LOADING, Exchange Done 00:02:53: %OSPF-5-ADJCHG: Process 1, Nbr 192.168.1.1 on Ethernet0/0 from LOADING to FULL, Loading Done ! Next R5 has costs of 50 and 64, respectively, on interfaces s0.2 and s0.1. R5# show ip ospf int brief Interface

PID

Area

IP Address/Mask

Cost

State Nbrs F/C

Se0.2

1

5

10.5.25.5/24

50

P2P

1/1

Se0.1

1

5

10.5.15.5/24

64

P2P

1/1

Et0

1

5

10.5.1.5/24

10

DR

0/0

! Below, S0.1’s cost was based on bandwidth of 64, using formula 108 / bandwidth, ! with bandwidth in bits/second. R5# sh int s 0.1 Serial0.1 is up, line protocol is up Hardware is HD64570 Internet address is 10.5.15.5/24 MTU 1500 bytes, BW 1544 Kbit, DLY 20000 usec, reliability 255/255, txload 1/255, rxload 1/255 Encapsulation FRAME-RELAY Last clearing of "show interface" counters never ! Next, R2’s interface costs are shown, including the minimum cost 1 on fa0/0. R2# sho ip ospf int brief Interface

PID

Area

IP Address/Mask

Cost

State Nbrs F/C

Fa0/0

2

0

10.1.1.2/24

1

BDR

3/3

Se0/0.5

2

5

10.5.25.2/24

64

P2P

1/1

! Below, R2 changes its reference bandwidth from the default of 100 Mbps to ! 10,000 Mbps. That in turn changes R2’s calculated cost values to be 100 times ! larger than before. Note that IOS allows this setting to differ on the routers, ! but recommends against it. R2# conf t Enter configuration commands, one per line.

End with CNTL/Z.

R2(config)# router ospf 2 R2(config-router)# auto-cost reference-bandwidth 10000 % OSPF: Reference bandwidth is changed. Please ensure reference bandwidth is consistent across all routers. R2# show ip ospf int brief Interface

PID

Area

IP Address/Mask

Cost

State Nbrs F/C

Fa0/0

2

0

10.1.1.2/24

100

BDR

3/3

Se0/0.5

2

5

10.5.25.2/24

6476

P2P

1/1

292

Chapter 8: OSPF

While Examples 8-8 and 8-9 show some details, the following list summarizes how IOS chooses OSPF interface costs: 1.

Set the cost per neighbor using the neighbor neighbor cost value command. (This is valid only on OSPF network types that allow neighbor commands.)

2.

Set the cost per interface using the ip ospf cost value interface subcommand.

3.

Allow cost to default based on interface bandwidth and the OSPF Reference Bandwidth (Ref-BW) (default 108). The formula is Ref-BW / bandwidth (bps).

4.

Default based on bandwidth, but change Ref-BW using the command auto-cost referencebandwidth value command within the OSPF process.

The only slightly tricky part of the cost calculation math is to keep the units straight, because the IOS interface bandwidth is kept in kbps, and the auto-cost reference-bandwidth command’s units are Mbps. For instance, on R5 in Example 8-9, the cost is calculated as 100 Mbps divided by 1544 kbps, where 1544 kbps is equal to 1.544 Mbps. The result is rounded down to the nearest integer, 64 in this case. On R2’s fa0/0, the bandwidth is 100,000 kbps, or 100 Mbps, making the calculation yield a cost of 1. After changing the reference bandwidth to 10,000, which means 10,000 Mbps, R2’s calculated costs were 100 times larger. NOTE When choosing the best routes to reach a subnet, OSPF also considers whether a route is an intra-area route, inter-area route, E1 route, or E2 route. OSPF prefers intra-area over all the rest, then interarea, then E1, and finally E2 routes. Under normal circumstances, routes to a single subnet should all be the same type; however, it is possible to have multiple route paths to reach a single subnet in the OSPF SPF tree, but with some of these routes being a different type. Example 9-7 in Chapter 9 demonstrates this.

Alternatives to the OSPF Network Command As of Cisco IOS Software Release 12.3(11)T, OSPF configuration can completely omit the network command, instead relying on the ip ospf process-id area area-id interface subcommand. This new command enables OSPF on the interface and selects the area. For instance, on R3 in Example 8-8, the network 10.3.0.0 0.0.255.255 area 3 command could have been deleted and replaced with the ip ospf 1 area 3 command under S0/0.1 and e0/0. The network and ip ospf area commands have some minor differences when secondary IP addresses are used. With the network command, OSPF advertises stub networks for any secondary IP subnets that are matched by the command. (“Secondary subnet” is jargon that refers to the subnet in which a secondary IP address resides.) The ip ospf area interface subcommand causes any and all secondary subnets on the interface to be advertised as stub networks—unless the optional secondaries none parameter is included at the end of the command.

OSPF Configuration

293

OSPF Filtering Intra-routing–protocol filtering presents some special challenges with link-state routing protocols like OSPF. Link-state protocols do not advertise routes—they advertise topology information. Also, SPF loop prevention relies on each router in the same area having an identical copy of the LSDB for that area. Filtering could conceivably make the LSDBs differ on different routers, causing routing irregularities. IOS supports three variations of what could loosely be categorized as OSPF route filtering. These three major types of OSPF filtering are as follows: ■

Filtering routes, not LSAs—Using the distribute-list in command, a router can filter the routes its SPF process is attempting to add to its routing table, without affecting the LSDB.

■

ABR type 3 LSA filtering—A process of preventing an ABR from creating particular type 3 summary LSAs.

■

Using the area range no-advertise option—Another process to prevent an ABR from creating specific type 3 summary LSAs.

Each of these three topics is discussed in sequence in the next few sections. Filtering Routes Using the distribute-list Command For RIP and EIGRP, the distribute-list command can be used to filter incoming and outgoing routing updates. The process is straightforward, with the distribute-list command referring to ACLs or prefix lists. With OSPF, the distribute-list command filters what ends up in the IP routing table, and on only the router on which the distribute-list command is configured. NOTE The redistribute command, when used for route distribution between OSPF and other routing protocols, does control what enters and leaves the LSDB. Chapter 9 covers more on route redistribution. The following rules govern the use of distribute lists for OSPF, when not used for route redistribution with other routing protocols: ■

Distribute lists can be used only for inbound filtering, because filtering any outbound OSPF information would mean filtering LSAs, not routes.

■

The inbound logic does not filter inbound LSAs; it instead filters the routes that SPF chooses to add to that one router’s routing table.

■

If the distribute list includes the incoming interface parameter, the incoming interface is checked as if it were the outgoing interface of the route.

That last bullet could use a little clarification. For example, if R2 learns routes via RIP or EIGRP updates that enter R2’s s0/0 interface, those routes typically use R2’s s0/0 interface as the outgoing

294

Chapter 8: OSPF

interface of the routes. The OSPF LSAs may have been flooded into a router on several interfaces, so an OSPF router checks the outgoing interface of the route as if it had learned about the routes via updates coming in that interface. Example 8-10 shows an example of two distribute lists on R5 from Figure 8-6. The example shows two options to achieve the same goal. In this case, R5 will filter the route to 10.4.8.0/24 via R5’s S0.2 subinterface (to R2). Later, it uses a route map to achieve the same result. Example 8-10

Filtering Routes with OSPF distribute-list Commands on R5

! R5 has a route to 10.4.8.0/24 through R2 (10.5.25.2, s0.2) R5# sh ip route ospf | incl 10.4.8.0 O IA

10.4.8.0/24 [110/1623] via 10.5.25.2, 00:00:28, Serial0.2

! Next, the distribute-list command refers to a prefix list that permits 10.4.8.0 ! /24. ip prefix-list prefix-9-4-8-0 seq 5 deny 10.4.8.0/24 i p p r e f i x - l i s t p r e f i x - 9 - 4 - 8 - 0 s e q 1 0 p e r m i t 0 . 0 . 0. 0 / 0 l e 3 2 ! Router ospf 1 d i s t r i b u t e - l i s t p r e f i x p re f i x - 9 - 4 - 8 - 0 i n S e r i a l 0 . 2 ! Below, note that R5’s route through R2 is gone, but the LSDB is unchanged! R5# sh ip route ospf | incl 10.4.8.0 ! Not shown: the earlier distribute-list command is removed. ! Below, note that the distribute-list command with the route-map option does not ! have an option to refer to an interface, so the route map itself has been ! configured to refer to the advertising router’s RID (2.2.2.2). Router ospf 1 distribute-list route-map lose-9-4-8-0 in ! Next, ACL 48 matches the 10.4.8.0/24 prefix, with ACL 51 matching R2’s RID. a c c e s s - l i s t 4 8 p e r m i t 1 0 .4 . 8 . 0 a c c e s s - l i s t 5 1 p e r m it 2 . 2 . 2 . 2 ! Below, the route map matches the prefix (based on ACL 48) and the advertising ! RID (ACL 51, matching R2’s 2.2.2.2 RID). Clause 20 permits all other prefixes. route-map lose-9-4-8-0 deny 10 match i p address 48 match ip ro ute-source 51 route-map lose-9-4-8-0 permit 20 ! Above, note the same results as the previous distribute list. R5# sh ip route ospf | incl 10.4.8.0

Example 8-10 shows only two ways to filter the routes. The distribute-list route-map option, allows a much greater variety of matching parameters, and much more detailed logic with route maps. For instance, this example showed matching a prefix as well as the RID that advertised the LSA to R5, namely 2.2.2.2 (R2). Refer to Chapter 10 for a more complete review of route maps and the match command.

OSPF Configuration

295

NOTE Some earlier IOS releases allowed the router to not only filter the route as shown in Example 8-7, but to replace the route with the next best route. Testing at 12.4 and beyond shows the behavior as shown in the example, with IOS simply not adding the route to the IP routing table.

OSPF ABR LSA Type 3 Filtering ABRs do not forward type 1 and 2 LSAs from one area into another, but instead create type 3 LSAs for each subnet defined in the type 1 and 2 LSAs. Type 3 LSAs do not contain detailed information about the topology of the originating area; instead, each type 3 LSA represents a subnet, and a cost from the ABR to that subnet. The earlier section “LSA Type 3 and Inter-Area Costs” covers the details and provides an example. The OSPF ABR type 3 LSA filtering feature allows an ABR to filter type 3 LSAs at the point where the LSAs would normally be created. By filtering at the ABR, before the type 3 LSA is injected into another area, the requirement for identical LSDBs inside the area can be met, while still filtering LSAs. To configure type 3 LSA filtering, you use the area number filter-list prefix name in | out command under router ospf. The referenced prefix list is used to match the subnets and masks to be filtered. The area number and the in | out option of the area filter-list command work together, as follows: ■

When in is configured, IOS filters prefixes going into the configured area.

■

When out is configured, IOS filters prefixes coming out of the configured area.

Example 8-11 should clarify the basic operation. ABR R1 will use two alternative area filter-list commands, both to filter subnet 10.3.2.0/23, the subnet that exists between R3 and R33 in Figure 8-6. Remember that R1 is connected to areas 0, 3, 4, and 5. The first area filter-list command shows filtering the LSA as it goes out of area 3; as a result, R2 will not inject the LSA into any of the other areas. The second case shows the same subnet being filtered going into area 0, meaning that the type 3 LSA for that subnet still gets into the area 4 and 5 LSDBs. Example 8-11

Type 3 LSA Filtering on R1 with the area filter-list Command

! The command lists three lines of extracted output. One line is for the ! type 3 LSA in area 0, one is for area 4, and one is for area 5. R1# show ip ospf data summary | include 10.3.2.0 Link State ID: 10.3.2.0 (summary Network Number) Link State ID: 10.3.2.0 (summary Network Number) Link State ID: 10.3.2.0 (summary Network Number) ! Below, the two-line prefix list denies subnet 10.3.2.0/23, and then permits ! all others. ip prefix-list filter-type3-9-3-2-0 seq 5 deny 10.3.2.0/23 i p p r e f i x - l i s t f il t e r - t y p e 3 - 9 - 3 - 2 - 0 s e q 1 0 p e r m i t 0 . 0 . 0 . 0 / 0 l e 3 2 Next, the area filter-list command filters type 3 LSAs going out of area 3. R1# conf t

continues

296

Chapter 8: OSPF

Example 8-11

Type 3 LSA Filtering on R1 with the area filter-list Command (Continued)

Enter configuration commands, one per line.

End with CNTL/Z.

R1(config)# router ospf 1 R1(config-router)# area 3 filter-list prefix filter-type3-9-3-2-0 out R1(config-router)# ^Z ! Below, R1 no longer has any type 3 LSAs, in areas 0, 4, and 5. For ! comparison, this command was issued a few commands ago, listing 1 line ! of output for each of the other 3 areas besides area 3. R1# show ip ospf data | include 10.3.2.0 ! Below, the previous area filter-list command is replaced by the next command ! below, which filters type 3 LSAs going into area 0, with the same prefix list. area 0 fi lter-list prefix filter-type3-9-3- 2-0 in ! Next, only 2 type 3 LSAs for 10.3.2.0 are shown—the ones in areas 4 and 5. R1# show ip ospf data | include 10.3.2.0 Link State ID: 10.3.2.0 (summary Network Number) Link State ID: 10.3.2.0 (summary Network Number) ! Below, the configuration for filtering type 3 LSAs with the area range command, ! which is explained following this example. The existing area filter-list ! commands from earlier in this chapter have been removed at this point. R1(config-router)# area 3 range 10.3.2.0 255.255.254.0 not-advertise R1# show ip ospf data summary | include 10.3.2.0 R1#

Filtering Type 3 LSAs with the area range Command The third method to filter OSPF routes is to filter type 3 LSAs at an ABR using the area range command. The area range command performs route summarization at ABRs, telling a router to cease advertising smaller subnets in a particular address range, instead creating a single type 3 LSA whose address and prefix encompass the smaller subnets. When the area range command includes the not-advertise keyword, not only are the smaller component subnets not advertised as type 3 LSAs, the summary route is not advertised as a type 3 LSA either. As a result, this command has the same effect as the area filter-list command with the out keyword, filtering the LSA from going out to any other areas. An example area range command is shown at the end of Example 8-11.

Virtual Link Configuration OSPF requires that each non-backbone area be connected to the backbone area (area 0). OSPF also requires that the routers in each area have a contiguous intra-area path to the other routers in the same area, because without that path, LSA flooding inside the area would fail. However, in some designs, meeting these requirements might be a challenge. You can use OSPF virtual links to overcome these problems. For instance, in the top part of Figure 8-12, area 33 connects only to area 3, and not to area 0.

OSPF Configuration Figure 8-12

297

The Need for Virtual Links Virtual Link – Transit Area 3

R33

Area 33

R3

Area 4’s Failed Links have a on them

Area 3

R1

R5

Area 0 R6

R4

Area 4

R7

S1 Virtual Link

R8

One straightforward solution to area 33’s lack of connection to the backbone area would be to combine areas 3 and 33 into a single area, but OSPF virtual links could solve the problem as well. An OSPF virtual link allows a pair of routers to tunnel OSPF packets inside IP packets, across the IP network, to some other router that is not on the same data link. A virtual link between R3 and R1 gives area 33 a connection to area 0. Also note that R3 becomes an ABR, with a full copy of area 0’s LSDB entries. While the top part of Figure 8-10 simply shows a possibly poor OSPF area design, the lower part shows what could happen just because of a particular set of link failures. The figure shows several failed links that result in a partitioned area 4. As a result of the failures, R7 and R8 have no area 4 links connecting to the other three routers in area 4. A virtual link can be used to connect R4 and R8—the requirement being that both R4 and R8 connect to a common and working area— recombining the partitions through the virtual link. (A better solution than the virtual link in this particular topology might be to trunk on R4 and R8, create a small subnet through the LAN switch, and put it in area 4.) Example 8-12 demonstrates a virtual link configuration between R33 and R1, as shown in Figure 8-12. Note that the virtual link cannot pass through a transit area that is a stubby area, so area 3 has been changed to no longer be a stubby area. Example 8-12

Virtual Link Between R3 and R1

! R1 has not learned subnet 10.3.2.0 yet, because area 33 has no link to area 0. R1# show ip route ospf | incl 10.3.2.0 R1# ! the area virtual link commands point to the other router’s RID, and the ! transit area over which the virtual link exists—area 3 in this case. Note that

continues

298

Chapter 8: OSPF

Virtual Link Between R3 and R1 (Continued)

Example 8-12

! timers can be set on the area virtual-link command, as well as authentication. ! It is important when authenticating virtual links to remember that ! the virtual links themselves are in area 0. ! R1 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! router ospf 1 a r e a 3 v i rt u a l - l i n k 3 . 3 . 3 . 3 ! R3 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! ro uter ospf 1 area 3 virtual-link 1.1.1.1 ! Below, the status of the virtual link is listed. R1# show ip ospf virtual-links Virtual Link OSPF_VL0 to router 3.3.3.3 is up Run as demand circuit DoNotAge LSA allowed. Transit area 3, via interface Serial0/0.3, Cost of using 64 Transmit Delay is 1 sec, State POINT_TO_POINT, Timer intervals configured, Hello 10, Dead 40, Wait 40, Retransmit 5 Hello due in 00:00:02 Adjacency State FULL (Hello suppressed) Index 3/6, retransmission queue length 0, number of retransmission 1 First 0x0(0)/0x0(0) Next 0x0(0)/0x0(0) Last retransmission scan length is 1, maximum is 1 Last retransmission scan time is 0 msec, maximum is 0 msec ! Because R1 and R3 are also sharing the same link, there is a neighbor ! relationship in area 3 that has been seen in the other examples, listed off ! interface s0/0.3. The new virtual link neighbor relationship is shown as well, ! with interface VL0 listed. R1# show ip ospf nei ! Lines omitted for brevity Neighbor ID

Pri

State

3.3.3.3

0

FULL/ —

3.3.3.3

0

FULL/ —

Dead Time — 00:00:10

Address

Interface

10.3.13.3

OSPF_VL0

10.3.13.3

Serial0/0.3

! Below, subnet 10.3.2.0/23, now in area 33, is learned by R1 over the Vlink. R1# show ip route ospf | incl 10.3.2.0 O IA

10.3.2.0/23 [110/75] via 10.3.13.3, 00:00:10, Serial0/0.3

Configuring OSPF Authentication One of the keys to keeping OSPF authentication configuration straight is to remember that it differs significantly with RIPv2 and EIGRP, although some of the concepts are very similar. The basic rules for configuring OSPF authentication are as follows: ■

Three types are available: type 0 (none), type 1 (clear text), and type 2 (MD5).

■

Authentication is enabled per interface using the ip ospf authentication interface subcommand.

OSPF Configuration

299

■

The default authentication is type 0 (no authentication).

■

The default can be redefined using the area authentication subcommand under router ospf.

■

The keys are configured as interface subcommands.

■

Multiple keys are allowed per interface; if configured, OSPF sends multiple copies of each message, one for each key.

Table 8-7 lists the three OSPF authentication types, along with the commands to enable each type, and the commands to define the authentication keys. Note that the three authentication types can be seen in the messages generated by the debug ip ospf adj command. Table 8-7

OSPF Authentication Types Type

Meaning

Enabling Interface Subcommand

Authentication Key Configuration Interface Subcommand

0

None

ip ospf authentication null

—

1

Clear text

ip ospf authentication

ip ospf authentication-key key-value

2

MD5

ip ospf authentication message-digest

ip ospf message-digest-key key-number md5 key-value

Example 8-13 (again based on Figure 8-6) shows examples of type 1 and type 2 authentication configuration routers R1 and R2. (Note that S1 and S2 have been shut down for this example, but they would need the same configuration as shown on R1 and R2.) In this example, both R1 and R2 use their fa0/0 interfaces, so their authentication configuration will be identical. As such, the example shows only the configuration on R1. Example 8-13

OSPF Authentication Using Only Interface Subcommands

! The two ip ospf commands are the same on R1 and R2. The first enables ! type 1 authentication, and the other defines the simple text key. interface FastE thernet0/0 ip ospf authentication ip o s p f a u t h e n t i c a t i o n - k e y k e y - t 1 ! Below, the neighbor relationship formed, proving that authentication works. R1# show ip ospf neighbor fa 0/0 Neighbor ID 2.2.2.2

Pri 1

State

Dead Time

Address

Interface

FULL/BDR

00:00:37

10.1.1.2

FastEthernet0/0

! Next, each interface’s OSPF authentication type can be seen in the last line ! or two in the output of the show ip ospf interface command. R1# show ip ospf int fa 0/0 ! Lines omitted for brevity Simple password authentication enabled ! Below, both R1 and R2 change to use type 2 authentication. Note that the key ! must be defined with the ip ospf message-digest-key interface subcommand. Key

continues

300

Chapter 8: OSPF

Example 8-13

OSPF Authentication Using Only Interface Subcommands

! chains cannot be used. interfac e FastEthernet0/ 0 ip osp f authentication message-digest ip ospf message-digest-key 1 md5 key-t2 ! Below, the command confirms type 2 (MD5) authentication, key number 1. R1# show ip ospf int fa 0/0 | begin auth ! Lines omitted for brevity Message digest authentication enabled Youngest key id is 1

Example 8-13 shows two working examples of OSPF authentication, neither of which uses the area number authentication under router ospf. Some texts imply that the area authentication command is required—in fact, it was required prior to Cisco IOS Software Release 12.0. In later IOS releases, the area authentication command simply tells the router to change that router’s default OSPF authentication type for all interfaces in that area. Table 8-8 summarizes the effects and syntax of the area authentication router subcommand. Table 8-8

Effect of the area authentication Command on OSPF Interface Authentication Settings area authentication Command

Interfaces in That Area Default to Use…

Type 0

area num authentication

Type 1

area num authentication message-digest

Type 2

The keys themselves are kept in clear text in the configuration, unless you add the service password-encryption global command to the configuration. The last piece of authentication configuration relates to OSPF virtual links. Because virtual links have no underlying interface on which to configure authentication, authentication is configured on the area virtual-link command itself. Table 8-9 shows the variations of the command options for configuring authentication on virtual links. Note that beyond the base area number virtual-link rid command, the parameters use similar keywords as compared with the equivalent interface subcommands. Table 8-9

Configuring OSPF Authentication on Virtual Links Type

Command Syntax for Virtual Links

0

area num virtual-link router-id authentication null

1

area num virtual-link router-id authentication authentication-key key-value

2

area num virtual-link router-id authentication message-digest message-digest-key key-num md5 key-value

OSPF Configuration

301

NOTE OSPF authentication is a good place for tricky CCIE lab questions—ones that can be solved in a few minutes if you know all the intricacies.

OSPF Stub Router Configuration Defined in RFC 3137, and first supported in Cisco IOS Software Release 12.2(4)T, the OSPF stub router feature—not to be confused with stubby areas—allows a router to either temporarily or permanently be prevented from becoming a transit router. In this context, a transit router is simply one to which packets are forwarded, with the expectation that the transit router will forward the packet to yet another router. Conversely, non-transit routers only forward packets to and from locally attached subnets. Figure 8-13 shows one typical case in which a stub router might be useful. Figure 8-13

OSPF Stub Router 2

ASBR1 Come Up: OSPF Convergence: 1 minute BGP Convergence: 5 minutes

BGP

ABR1

ASBR1 Internet

Default Routes BGP

ASBR2

ABR2

1

OSPF and BGP Already Converged

Both ASBR1 and ASBR2 advertise defaults into the network, expecting to have the capability to route to the Internet through BGP-learned routes. In this case, ASBR2 is already up, fully converged. However, if ASBR1 reloads, when it comes back up, OSPF is likely to converge faster than BGP. As a result, ASBR1 will advertise its default route, and OSPF routers may send packets to ASBR1, but ASBR1 will end up discarding the packets until BGP converges. Using the stub router feature on the ASBRs solves the problem by making them advertise infinite metric routes (cost 16,777,215) for any transit routes—either for a configured time period or until BGP convergence is complete. To do so, under router ospf, the ASBRs would use either the maxmetric router-lsa on-startup announce-time command or the max-metric router-lsa on-startup wait-for-bgp command. With the first version, the actual time period (in seconds) can be set. With the second, OSPF waits until BGP signals that convergence is complete or until 10 minutes pass, whichever comes first.

302

Chapter 8: OSPF

Foundation Summary This section lists additional details and facts to round out the coverage of the topics in this chapter. Unlike most of the Cisco Press Exam Certification Guides, this “Foundation Summary” does not repeat information presented in the “Foundation Topics” section of the chapter. Please take the time to read and study the details in the “Foundation Topics” section of the chapter, as well as review items noted with a Key Topic icon. Table 8-10 lists some of the key protocols regarding OSPF. Table 8-10

Protocols and Corresponding Standards for Chapter 8 Name

Standard

OSPF Version 2

RFC 2328

The OSPF Opaque LSA Option

RFC 5250

The OSPF Not-So-Stubby Area (NSSA) Option

RFC 3101

OSPF Stub Router Advertisement

RFC 3137

Traffic Engineering (TE) Extensions to OSPF Version 2

RFC 3630

Graceful OSPF Restart

RFC 3623

Table 8-11 lists some of the most popular IOS commands related to the topics in this chapter. Also, refer to Tables 8-7 through 8-9 for references to OSPF authentication commands. Table 8-11

Command Reference for Chapter 8 Command

Command Mode and Description

router ospf process-id

Global config; puts user in OSPF configuration mode for that PID.

network ip-address [wildcard-mask] area area

OSPF config mode; defines matching parameters, compared to interface IP addresses, to pick interfaces on which to enable OSPF.

ip ospf process-id area area-id [secondaries none]

Interface config mode; alternative to the network command for enabling OSPF on an interface.

neighbor ip-address [priority number] [poll-interval seconds] [cost number] [database-filter all]

OSPF config mode; used when neighbors must be defined, it identifies the neighbor’s IP address, priority, cost, and poll interval.

auto-cost reference-bandwidth ref-bw

OSPF config mode; changes the numerator in the formula to calculate interface cost for all OSPF interfaces on that router.

Foundation Summary

Table 8-11

303

Command Reference for Chapter 8 Command

Command Mode and Description

router-id ip-address

OSPF config mode; statically sets the router ID.

ospf log-neighbor-changes [detail]

EIGRP subcommand; displays log messages when neighbor status changes. On by default.

passive-interface [default] {interface-type interface-number}

OSPF config mode; causes OSPF to stop sending Hellos on the specified interface. OSPF will still advertise the subnet as a stub network.

area area-id stub [no-summary]

OSPF config mode; sets the area type to stub or totally stubby.

area area-id nssa [no-redistribution] [default-information-originate [metric] [metric-type]] [no-summary]

OSPF config mode; sets the area type to NSSA or totally NSSA.

area area-id default-cost cost

OSPF config mode; sets the cost of default route created by ABRs and sent into stubby areas.

area area-id nssa translate type7 suppress-fa

OSPF config mode; sets an NSSA ABR to set the forwarding address to 0.0.0.0 for the type 5 LSAs it translates from type 7.

area area-id range ip-address mask [advertise | not-advertise] [cost cost]

OSPF config mode; summarizes routes into a larger prefix at ABRs. Optionally filters type 3 LSAs (not-advertise option).

area {area-id} filter-list prefix {prefixlist-name in | out}

OSPF config mode; filters type 3 LSA creation at ABR.

distribute-list [ACL] | [route-map map-tag] in [int-type | int-number]

OSPF config mode; defines ACL or prefix list to filter what OSPF puts into the routing table.

area area-id virtual-link router-id [authentication [message-digest | null]] [hello-interval seconds] [retransmit-interval seconds] [transmit-delay seconds] [deadinterval seconds] [[authenticationkey key] | [message-digest-key key-id md5 key]]

OSPF config mode; creates a virtual link, with typical interface configuration settings to overcome fact that the link is virtual.

ip ospf hello-interval seconds

Interface subcommand; sets the interval for periodic Hellos.

ip ospf dead-interval {seconds | minimal hello-multiplier multiplier}

Interface subcommand; defines the dead interval, or optionally the minimal dead interval of 1 second.

ip ospf name-lookup

Global command; causes the router to use DNS to correlate RIDs to host names for show command output. continues

304

Chapter 8: OSPF

Table 8-11

Command Reference for Chapter 8 Command

Command Mode and Description

ip ospf cost interface-cost

Interface subcommand; sets the cost.

ip ospf mtu-ignore

Interface subcommand; tells the router to ignore the check for equal MTUs that occurs when sending DD packets.

ip ospf network {broadcast | nonbroadcast | {point-to-multipoint [non-broadcast] | point-to-point}}

Interface subcommand; sets the OSPF network type on an interface.

ip ospf priority number-value

Interface subcommand; sets the OSPF priority on an interface.

ip ospf retransmit-interval seconds

Interface subcommand; sets the time between LSA transmissions for adjacencies belonging to an interface.

ip ospf transmit-delay seconds

Interface subcommand; defines the estimated time expected for the transmission of an LSU.

max-metric router-lsa [on-startup {announce-time | wait-for-bgp}]

OSPF config mode; configures a stub router, delaying the point at which it can become a transit router.

show ip ospf border-routers

User mode; displays hidden LSAs for ABRs and ASBRs.

show ip ospf [process-id [area-id]] database

User mode; has many options not shown here. Displays the OSPF LSDB.

show ip ospf neighbor [interface-type interface-number] [neighbor-id] [detail]

User mode; lists information about OSPF neighbors.

show ip ospf [process-id] summaryaddress

User mode; lists information about route summaries in OSPF.

show ip ospf virtual-links

User mode; displays status and info about virtual links.

show ip route ospf

User mode; displays all OSPF routes in the IP routing table.

show ip ospf interface [interface-type interface-number] [brief]

User mode; lists OSPF protocol timers and statistics per interface.

show ip ospf statistics [detail]

User mode; displays OSPF SPF calculation statistics.

clear ip ospf [pid] {process | redistribution | counters [neighbor [neighbor-interface] [neighbor-id]]}

Enable mode; restarts the OSPF process, clears redistributed routes, or clears OSPF counters.

debug ip ospf hello

Enable mode; displays messages regarding Hellos, including Hello parameter mismatches.

debug ip ospf adj

Enable mode; displays messages regarding adjacency changes.

show ip ospf interface [type number] [brief]

User mode; lists many interface settings.

Foundation Summary

305

Table 8-12 summarizes many OSPF timers and their meaning. Table 8-12

OSPF Timer Summary Timer

Meaning

MaxAge

The maximum time an LSA can be in a router’s LSDB, without receiving a newer copy of the LSA, before the LSA is removed. Default is 3600 seconds.

LSRefresh

The timer interval per LSA on which a router refloods an identical LSA, except for a 1-larger sequence number, to prevent the expiration of MaxAge. Default is 1800 seconds.

Hello

Per interface; time interval between Hellos. Default is 10 or 30 seconds, depending on interface type.

Dead

Per interface; time interval in which a Hello should be received from a neighbor. If not received, the neighbor is considered to have failed. Default is four times Hello.

Wait

Per interface; set to the same number as the dead interval. Defines the time a router will wait to get a Hello asserting a DR after reaching a 2WAY state with that neighbor.

Retransmission

Per interface; the time between sending an LSU, not receiving an acknowledgement, and then resending the LSU. Default is 5 seconds.

Inactivity

Countdown timer, per neighbor, used to detect when a neighbor has not been heard from for a complete dead interval. It starts equal to the dead interval, counts down, and is reset to be equal to the dead interval when each Hello is received.

Poll Interval

On NBMA networks, the period at which Hellos are sent to a neighbor when the neighbor is down. Default is 60 seconds.

Flood (Pacing)

Per interface; defines the interval between successive LSUs when flooding LSAs. Default is 33 ms.

Retransmission (Pacing)

Per interface; defines the interval between retransmitted packets as part of a single retransmission event. Default is 66 ms.

Lsa-group (Pacing)

Per OSPF process. LSA’s LSRefresh intervals time out independently. This timer improves LSU reflooding efficiency by waiting, collecting several LSAs whose LSRefresh timers expire, and flooding all these LSAs together. Default is 240 seconds.

306

Chapter 8: OSPF

Table 8-13 lists OSPF neighbor states and their meaning. Table 8-13

OSPF Neighbor States State

Meaning

Down

No Hellos have been received from this neighbor for more than the dead interval.

Attempt

This router is sending Hellos to a manually configured neighbor.

Init

A Hello has been received from the neighbor, but it did not have the router’s RID in it. This is a permanent state when Hello parameters do not match.

2WAY

A Hello has been received from the neighbor, and it has the router’s RID in it. This is a stable state for pairs of DROther neighbors.

ExStart

Currently negotiating the DD sequence numbers and master/slave logic used for DD packets.

Exchange

Finished negotiating, and currently exchanging DD packets.

Loading

All DD packets exchanged, and currently pulling the complete LSDB entries with LSU packets.

Full

Neighbors are adjacent (fully adjacent), and should have identical LSDB entries for the area in which the link resides. Routing table calculations begin.

Table 8-14 lists several key OSPF numeric values. Table 8-14

OSPF Numeric Ranges Setting

Range of Values

Single interface cost

1 to 65,535 (216 – 1)

Complete route cost

1 to 16,777,215 (224 – 1)

Infinite route cost

16,777,215 (224 – 1)

Reference bandwidth (units: Mbps)

1 to 4,294,967

OSPF PID

1 to 65,535 (216 – 1)

Memory Builders The CCIE Routing and Switching written exam, like all Cisco CCIE written exams, covers a fairly broad set of topics. This section provides some basic tools to help you exercise your memory about some of the broader topics covered in this chapter.

Further Reading

307

Fill In Key Tables from Memory Appendix G, “Key Tables for CCIE Study,” on the CD in the back of this book contains empty sets of some of the key summary tables in each chapter. Print Appendix G, refer to this chapter’s tables in it, and fill in the tables from memory. Refer to Appendix H, “Solutions for Key Tables for CCIE Study,” on the CD to check your answers.

Definitions Next, take a few moments to write down the definitions for the following terms: LSDB, Dijkstra, link-state routing protocol, LSA, LSU, DD, Hello, LSAck, RID, neighbor state, neighbor, adjacent, fully adjacent, 2-Way, 224.0.0.5, 224.0.0.6, area, stub area type, network type, external route, E1 route, E2 route, Hello timer, dead time/ interval, sequence number, DR, BDR, DROther, priority, LSA flooding, DR election, SPF calculation, partial SPF calculation, full SPF calculation, LSRefresh, hello time/ interval, Maxage, ABR, ASBR, internal router, backbone area, transit network, stub network, LSA type, stub area, NSSA, totally stubby area, totally NSSA area, virtual link, stub router, transit router, SPF algorithm, All OSPF DR Routers, All OSPF Routers, graceful restart Refer to the glossary to check your answers.

Further Reading Jeff Doyle’s Routing TCP/IP, Volume I, Second Edition—every word a must for CCIE Routing and Switching. Cisco OSPF Command and Configuration Handbook, by Dr. William Parkhurst, covers every OSPF-related command available in Cisco IOS at the time of that book’s publication, with examples of each one.

Blueprint topics covered in this chapter: This chapter covers the following subtopics from the Cisco CCIE Routing and Switching written exam blueprint. Refer to the full blueprint in Table I-1 in the Introduction for more details on the topics covered in each chapter and their context within the blueprint. ■

Manual Summarization and Autosummarization

■

Route Redistribution

■

Default Routing

■

Troubleshooting Complex Layer 3 Problems

CHAPTER

9

IGP Route Redistribution, Route Summarization, Default Routing, and Troubleshooting This chapter covers several topics related to the use of multiple IGP routing protocols. IGPs can use default routes to pull packets toward a small set of routers, with those routers having learned routes from some external source. IGPs can use route summarization with a single routing protocol, but it is often used at redistribution points between IGPs as well. Route redistribution by definition involves moving routes from one routing source to another. This chapter takes a look at each topic. New to the qualification exam blueprint are a number of troubleshooting topics. One of them, troubleshooting complex Layer 3 problems, is covered in this chapter. The goal is to provide you with a process and tools to troubleshoot these types of problems. For perspective, note that this chapter includes coverage of RIPv2 redistribution topics. Even though RIPv2 has been removed from the CCIE Routing and Switching qualifying exam blueprint, it is still possible that you might see exam questions on redistribution involving RIPv2. Therefore, this chapter includes coverage of that topic.

“Do I Know This Already?” Quiz Table 9-1 outlines the major headings in this chapter and the corresponding “Do I Know This Already?” quiz questions. Table 9-1

“Do I Know This Already?” Foundation Topics Section-to-Question Mapping Foundation Topics Section

Questions Covered in This Section

Route Maps, Prefix Lists, and Administrative Distance

1–2

Route Redistribution

3–6

Route Summarization

7

Default Routes

8

Troubleshooting Layer 3 Problems

9–10

Total Score

Score

310

Chapter 9: IGP Route Redistribution, Route Summarization, Default Routing, . . .

To best use this pre-chapter assessment, remember to score yourself strictly. You can find the answers in Appendix A, “Answers to the ‘Do I Know This Already?’ Quizzes.” 1.

2.

3.

A route map has several clauses. A route map’s first clause has a permit action configured. The match command for this clause refers to an ACL that matches route 10.1.1.0/24 with a permit action, and matches route 10.1.2.0/24 with a deny action. If this route map is used for route redistribution, which of the following are true? a.

The route map will attempt to redistribute 10.1.1.0/24.

b.

The question does not supply enough information to determine if 10.1.1.0/24 is redistributed.

c.

The route map will not attempt to redistribute 10.1.2.0/24.

d.

The question does not supply enough information to determine if 10.1.2.0/24 is redistributed.

Which of the following routes would be matched by this prefix list command: ip prefix-list fred permit 10.128.0.0/9 ge 20? a.

10.1.1.0 255.255.255.0

b.

10.127.1.0 255.255.255.0

c.

10.200.200.192 255.255.255.252

d.

10.128.0.0 255.255.240.0

e.

None of these answers is correct.

A router is using the configuration shown below to redistribute routes. This router has several working interfaces with IP addresses in network 10.0.0.0, and has learned some network 10 routes with EIGRP and some with OSPF. Which of the following is true about the redistribution configuration? router eigrp 1 network 10.0.0.0 redistribute ospf 2 ! router ospf 2 network 10.0.0.0 0.255.255.255 area 3 redistribute eigrp 1 subnets R1# show ip route 15.0.0.0 Routing entry for 15.0.0.0/24, 5 known subnets Attached (2 connections) Redistributing via eigrp 1 O E1 O E2

10.6.11.0 [110/84] via 10.1.6.6, 00:21:52, Serial0/0/0.6 10.6.12.0 [110/20] via 10.1.6.6, 00:21:52, Serial0/0/0.6

“Do I Know This Already?” Quiz

C O IA C

4.

311

10.1.6.0 is directly connected, Serial0/0/0.6 10.1.2.0 [110/65] via 10.1.1.5, 00:21:52, Serial0/0/0.5 10.1.1.0 is directly connected, Serial0/0/0.5

a.

EIGRP will not advertise any additional routes due to redistribution.

b.

OSPF will not advertise any additional routes due to redistribution.

c.

Routes redistributed into OSPF will be advertised as E1 routes.

d.

The redistribute ospf 2 command would be rejected due to missing parameters.

Examine the following router configuration and excerpt from its IP routing table. Which routes could be redistributed into OSPF? router eigrp 1 network 12.0.0.0 router ospf 2 redistribute eigrp 1 subnets network 13.0.0.0 0.255.255.255 area 3 An excerpt from the routing table is shown next: C 12.1.6.0 is directly connected, Serial0/0/0.6 D 12.0.0.0/8 [90/2172416] via 13.1.1.1, 00:01:30, Serial0/0/0.5 C 13.1.1.0 is directly connected, Serial0/0/0.5

5.

a.

12.1.6.0

b.

12.0.0.0

c.

13.1.1.0

d.

None of these answers is correct.

Two corporations merged. The network engineers decided to redistribute between one company’s EIGRP network and the other company’s OSPF network, using two mutually redistributing routers (R1 and R2) for redundancy. Assume that as many defaults as is possible are used for the redistribution configuration. Assume that one of the subnets in the OSPF domain is 10.1.1.0/24. Which of the following is true about a possible suboptimal route to 10.1.1.0/24 on R1—a route that sends packets through the EIGRP domain, and through R2 into the OSPF domain? a.

The suboptimal routes will occur unless the configuration filters routes at R1.

b.

R1’s administrative distance must be manipulated, such that OSPF routes have an administrative distance less than EIGRP’s default of 90.

c.

EIGRP prevents the suboptimal routes by default.

d.

Using route tags is the only way to prevent the suboptimal routes.

312

Chapter 9: IGP Route Redistribution, Route Summarization, Default Routing, . . .

6.

7.

8.

9.

Which of the following statements is true about the type of routes created when redistributing routes? a.

Routes redistributed into OSPF default to be external type 2.

b.

Routes redistributed into EIGRP default to external, but can be set to internal with a route map.

c.

Routes redistributed into RIP are external by default.

d.

Routes redistributed into OSPF by a router in an NSSA area default to be external type 1.

Which of the following is not true about route summarization? a.

The advertised summary is assigned the same metric as the lowest-metric component subnet.

b.

The router does not advertise the summary when its routing table does not have any of the component subnets.

c.

The router does not advertise the component subnets.

d.

Summarization, when used with redistribution, prevents all cases of suboptimal routes.

Which of the following is/are true regarding the default-information originate router subcommand? a.

It is not supported by EIGRP.

b.

It causes OSPF to advertise a default route, but only if a static route to 0.0.0.0/0 is in that router’s routing table.

c.

The always keyword on the default-information originate command, when used for OSPF, means OSPF will originate a default route even if no default route exists in its own IP routing table.

d.

None of the other answers is correct.

An EIGRP router is showing intermittent reachability to 172.30.8.32/27. Which command(s) reveals the source by which this prefix is being advertised to the local router? a.

show ip protocols

b.

show ip route eigrp

c.

show ip eigrp neighbor

d.

show ip eigrp topology 172.30.8.32

e.

show ip route 172.30.8.32

“Do I Know This Already?” Quiz

10.

You suspect that a routing loop exists in your network because a subnet is intermittently reachable. What is the most specific way to determine that a routing loop is the cause? a.

ping

b.

traceroute

c.

debug ip packet detail

d.

debug ip routing

e.

show ip protocols

313

314

Chapter 9: IGP Route Redistribution, Route Summarization, Default Routing, . . .

Foundation Topics

Route Maps, Prefix Lists, and Administrative Distance Route maps, IP prefix lists, and administrative distance (AD) must be well understood to do well with route redistribution topics on the CCIE Routing and Switching written exam. This section focuses on the tools themselves, followed by coverage of route redistribution.

Configuring Route Maps with the route-map Command Route maps provide programming logic similar to the If/Then/Else logic seen in other programming languages. A single route map has one or more route-map commands in it, and routers process route-map commands in sequential order based on sequence numbers. Each route-map command has underlying matching parameters, configured with the aptly named match command. (To match all packets, the route-map clause simply omits the match command.) Each route-map command also has one or more optional set commands that you can use to manipulate information— for instance, to set the metric for some redistributed routes. The general rules for route maps are as follows: ■

Each route-map command must have an explicitly configured name, with all commands that use the same name being part of the same route map.

■

Each route-map command has an action (permit or deny).

■

Each route-map command in the same route map has a unique sequence number, allowing deletion and insertion of single route-map commands.

■

When a route map is used for redistribution, the route map processes routes taken from the then-current routing table.

■

The route map is processed sequentially based on the sequence numbers.

■

Once a particular route is matched by the route map, it is not processed beyond that matching route-map command (specific to route redistribution).

■

When a route is matched in a route-map statement, if the route-map command has a permit parameter, the route is redistributed (specific to route redistribution).

■

When a route is matched in a route-map statement, if the route-map statement has a deny parameter, the route is not redistributed (specific to route redistribution).

Route Maps, Prefix Lists, and Administrative Distance

315

Route maps can be confusing at times, especially when using the deny option on the route-map command. To help make sure the logic is clear before getting into redistribution, Figure 9-1 shows a logic diagram for an example route map. (This example is contrived to demonstrate some nuances of route map logic; a better, more efficient route map could be created to achieve the same results.) In the figure, R1 has eight loopback interfaces configured to be in class A networks 32 through 39. Figure 9-1 shows how the contrived route-map picky would process the routes. Figure 9-1

Route Map Logic Example

A The set of routes to redistribute B 32, 33, 34, 35, 36, 37, 38, 39

B Routes for which no decision has been made yet

C The set of routes that are not Permit

redistributed Matched 32

route-map picky permit 10 match ip address 32 Not Matched

ACL 32: Permits Network 32 Default Deny Any at End

A 32

B 33, 34, 35, 36, 37, 38, 39

Permit Matched 37 ACL d-36-p-37: Denies Network 36 Permits Network 37 Default Deny Any at End

route-map picky permit 25 match ip address d-36-p-37 Not Matched

A 32, 37

B 33, 34, 35, 36, 38, 39

route-map picky deny 33 match ip address d-38-p39 ACL d-38-p-39: Denies Network 38 Permits Network 39 Default Deny Any at End

C null

Deny Matched 39

Not Matched

A 32, 37

B 33, 34, 35, 36, 38

route-map picky deny 40 match ip address 33 ACL 33: Permits Network 33 Default Deny Any at End

C null

C 39

Deny Matched 33

Not Matched

A 32, 37

B 34, 35, 36, 38

End of List: implied deny all A 32, 37

B Null

C 33, 39 Deny

C 33, 34, 35, 36, 38, 39

First, a few clarifications about the meaning of Figure 9-1 are in order. The top of the figure begins with the set of connected networks (32 through 39), labeled with a “B,” which is the set of routes still being considered for redistribution. Moving down the figure, four separate route-map commands sit inside this single route map. Each route-map clause (the clause includes the underlying match and set commands) in turn moves routes from the list of possible routes (“B”)

316

Chapter 9: IGP Route Redistribution, Route Summarization, Default Routing, . . .

to either the list of routes to redistribute (“A”) or the list to not redistribute (“C”). By the bottom of the figure, all routes will be noted as either to be redistributed or not to be redistributed. The route map chooses to redistribute a route only if the route-map command has a permit option; the only time a route-map clause chooses to not redistribute a route is when the clause has a deny option. Ignoring the matching logic for a moment, the first two route-map commands (sequence numbers 10 and 25) use the permit option. As a result of those clauses, routes are either added to the list of routes to redistribute (“A”) or left in the list of candidate routes (“B”). The third and fourth clauses (sequence numbers 33 and 40) use the deny option, so those clauses cause routes to be either added to the list of routes to not redistribute (“C”), or left in the list of candidate routes (“B”). In effect, once a route-map clause has matched a route, that route is flagged either as to be redistributed or as not to be redistributed, and the route is no longer processed by the route map. One point that can sometimes be confused is that if a route is denied by an ACL used by a match command, it does not mean that the route is prevented from being redistributed. For instance, the match ip address 32 command in clause 10 refers to ACL 32, which has one explicit access control entry (ACE) that matches network 32, with a permit action. Of course, ACL 32 has an implied deny all at the end, so ACL 32 permits network 32, and denies 33 through 39. However, denying networks 33 through 39 in the ACL does not mean that those routes are not redistributed—it simply means that those routes do not match route-map clause 10, so those routes are eligible for consideration by a later route-map clause. The following list summarizes the key points about route map logic when used for redistribution: ■

route-map commands with the permit option either cause a route to be redistributed or leave the route in the list of routes to be examined by the next route-map clause.

■

route-map commands with the deny option either filter the route or leave the route in the list of routes to be examined by the next route-map clause.

■

If a clause’s match commands use an ACL, an ACL match with the deny action does not cause the route to be filtered. Instead, it just means that route does not match that particular route-map clause.

■

The route-map command includes an implied deny all clause at the end; to configure a permit all, use the route-map command, with a permit action, but without a match command.

Route Map match Commands for Route Redistribution Route maps use the match command to define the fields and values used for matching the routes being processed. If more than one match command is configured in a single route-map clause, a route is matched only if all the match commands’ parameters match the route. The logic in each

Route Maps, Prefix Lists, and Administrative Distance

317

match command itself is relatively straightforward. Table 9-2 lists the match command options when used for IGP route redistribution. Table 9-2

match Command Options for IGP Redistribution match Command

Description

match interface interface-type interface-number [... interface-type interface-number]

Looks at outgoing interface of routes

*match ip address {[access-list-number | access-list-name] | prefix-list prefix-list-name}

Examines route prefix and prefix length

*match ip next-hop {access-list-number | access-list-name}

Examines route’s next-hop address

*match ip route-source {access-list-number | access-listname}

Matches advertising router’s IP address

match metric metric-value [+– deviation]

Matches route’s metric exactly, or optionally a range of metrics (plus/ minus the configured deviation.)

match route-type {internal | external [type-1 | type-2] | level-1 | level-2}

Matches route type

match tag tag-value [...tag-value]

Tag must have been set earlier

*Can reference multiple numbered and named ACLs on a single command.

Route Map set Commands for Route Redistribution When used for redistribution, route maps have an implied action—either to allow the route to be redistributed or to filter the route so that it is not redistributed. As described earlier in this chapter, that choice is implied by the permit or deny option on the route-map command. Route maps can also change information about the redistributed routes by using the set command. Table 9-3 lists the set command options when used for IGP route redistribution. Table 9-3

set Command Options for IGP Redistribution set Command

Description

set level {level-1 | level-2 | level-1-2 | stub-area | backbone}

Defines database(s) into which the route is redistributed

set metric metric-value

Sets the route’s metric for OSPF, RIP, and IS-IS

set metric bandwidth delay reliability loading mtu

Sets the IGRP/EIGRP route’s metric values

set metric-type {internal | external | type-1 | type-2}

Sets type of route for IS-IS and OSPF

set tag tag-value

Sets the unitless tag value in the route

318

Chapter 9: IGP Route Redistribution, Route Summarization, Default Routing, . . .

IP Prefix Lists IP prefix lists provide mechanisms to match two components of an IP route: ■

The route prefix (the subnet number)

■

The prefix length (the subnet mask)

The redistribute command cannot directly reference a prefix list, but a route map can refer to a prefix list by using the match command. A prefix list itself has similar characteristics to a route map. The list consists of one or more statements with the same text name. Each statement has a sequence number to allow deletion of individual commands, and insertion of commands into a particular sequence position. Each command has a permit or deny action—but because it is used only for matching packets, the permit or deny keyword just implies whether a route is matched (permit) or not (deny). The generic command syntax is as follows: ip prefix-list list-name [s seq seq-value] {d deny network/length | permit network/ / ge ge-value] [l le le-value] length}[g

The sometimes tricky and interesting part of working with prefix lists is that the meaning of the network/length, ge-value, and le-value parameters changes depending on the syntax. The network/ length parameters define the values to use to match the route prefix. For example, a network/length of 10.0.0.0/8 means “any route that begins with a 10 in the first octet.” The ge and le options are used for comparison to the prefix length—in other words, to the number of binary 1s in the subnet mask. For instance, ge 20 le 22 matches only routes whose masks are /20, /21, or /22. So, prefix list logic can be summarized into a two-step comparison process for each route: 1.

The route’s prefix must be within the range of addresses implied by the prefix-list command’s network/length parameters.

2.

The route’s prefix length must match the range of prefixes implied by the prefix-list command.

The potentially tricky part of the logic relates to knowing the range of prefix lengths checked by this logic. The range is defined by the ge-value and le-value parameters, which stand for greaterthan-or-equal-to and less-than-or-equal-to. Table 9-4 formalizes the logic, including the default values for ge-value and le-value. In the table, note that conf-length refers to the prefix length configured in the network/prefix (required) parameter, and route-length refers to the prefix length of a route being examined by the prefix list.

Route Maps, Prefix Lists, and Administrative Distance

Table 9-4

319

LE and GE Parameters on IP Prefix List, and the Implied Range of Prefix Lengths Prefix List Parameters

Range of Prefix Lengths

Neither

conf-length = route-length

Only le

conf-length